Blind and Visually Impaired Community
↑
7
↓
Hello, I am one of the annoying college assistive device makers wondering if I should stop (self.Blind)
submitted by SireSmiggle
Hello all!
Before I begin talking, I have already seen the README for assistive device makers posted by Rethunker. I found it about 7 months ago, and I have read a few of the biographies linked there and I am currently reading the 2nd textbook about O&M. Thank you Rethunker, it saved me quite a bit of trouble hunting down those resources myself.
I am a freshman in college and I have been working at my local sight center for 2 years now. I work part time during the school year and full time during the summer. I try to develop any odd ended tech to try and solve some issues that our VIs may be having. My most recent project is an app that uses RFID strips to replace braille labels and bump dots on pill bottles. We had more than one issue where some of our members were taking the wrong pills (Which is pretty dangerous), even with braille labels. All one has to do is point the phone at the RFID strip, and the phone will speak the text encoded in the strip. We expanded on that idea, and replaced some of our braille signs and labels around the center with long range RFID strips that phones can read from about a three meters away. It has been working out fairly well.
So... what got me into assistive devices was actually one of those "echolocation bands" that people have been dissing (Rightfully so in a lot of cases). I started working on a navigation hat about 4 years ago, and have since continued iterating it till today. It's changed significantly, but the premise remains about the same: gathering distance and position data all around the wearer using a bunch of sensors, and trying to communicate the data using haptic feedback.
I have asked many of the people at my local sight center about whether it is useful or not, and the majority of them said it would not be. However, the majority of them also do not require canes, dogs, or assistive devices to navigate in general because they still have enough sight remaining to navigate safely. (But not enough to read or use a computer without some assistance)
My question is: Would something like the device I am working on ever be useful to anyone here?
I am reluctant to slam the axe down on my little hat device because I have spent quite some time and multiple thousands of USD in personal funds to develop it, but I also understand that there's no point in continuing work on it if it is just a waste of my time and yours.
tl;dr: Are the echolocation/navigation devices *ever* usable to anyone with visual impairment?
Before I begin talking, I have already seen the README for assistive device makers posted by Rethunker. I found it about 7 months ago, and I have read a few of the biographies linked there and I am currently reading the 2nd textbook about O&M. Thank you Rethunker, it saved me quite a bit of trouble hunting down those resources myself.
I am a freshman in college and I have been working at my local sight center for 2 years now. I work part time during the school year and full time during the summer. I try to develop any odd ended tech to try and solve some issues that our VIs may be having. My most recent project is an app that uses RFID strips to replace braille labels and bump dots on pill bottles. We had more than one issue where some of our members were taking the wrong pills (Which is pretty dangerous), even with braille labels. All one has to do is point the phone at the RFID strip, and the phone will speak the text encoded in the strip. We expanded on that idea, and replaced some of our braille signs and labels around the center with long range RFID strips that phones can read from about a three meters away. It has been working out fairly well.
So... what got me into assistive devices was actually one of those "echolocation bands" that people have been dissing (Rightfully so in a lot of cases). I started working on a navigation hat about 4 years ago, and have since continued iterating it till today. It's changed significantly, but the premise remains about the same: gathering distance and position data all around the wearer using a bunch of sensors, and trying to communicate the data using haptic feedback.
I have asked many of the people at my local sight center about whether it is useful or not, and the majority of them said it would not be. However, the majority of them also do not require canes, dogs, or assistive devices to navigate in general because they still have enough sight remaining to navigate safely. (But not enough to read or use a computer without some assistance)
My question is: Would something like the device I am working on ever be useful to anyone here?
I am reluctant to slam the axe down on my little hat device because I have spent quite some time and multiple thousands of USD in personal funds to develop it, but I also understand that there's no point in continuing work on it if it is just a waste of my time and yours.
tl;dr: Are the echolocation/navigation devices *ever* usable to anyone with visual impairment?
codeplaysleep 3 points 2y ago
To be honest, I don't think it would have broad appeal.
But pill bottles with built-in NFC chips in the lids (or elsewhere on the bottle/label) and hardware/software so that the pharmacy could quickly transfer prescription data to the chip that you could read with an app on your phone? That's great! Run with that one.
But pill bottles with built-in NFC chips in the lids (or elsewhere on the bottle/label) and hardware/software so that the pharmacy could quickly transfer prescription data to the chip that you could read with an app on your phone? That's great! Run with that one.
SireSmiggle [OP] 2 points 2y ago
That's actually a great idea. I didn't even think of putting it in a pharmacy. I was just making a little custom device that makes it easy for people to make their own custom RFID stickers so they can label their own pill bottles and other items. However, just having it pre-integrated into the pharmacy sounds awesome!
I'll talk to my mentor about it. Our sight center works in conjunction with our local hospital, so we can maybe partner with them to get RFID labels on medicine in the pharmacy side.
I'll talk to my mentor about it. Our sight center works in conjunction with our local hospital, so we can maybe partner with them to get RFID labels on medicine in the pharmacy side.
LadyAlleta 3 points 2y ago
I'm sure that someone somewhere might need it. But to be honest, I'm lazy. So even though I can't see in sunlight at all (except the color yellow) I probably wouldn't go through the hassle of using it bc I'm running late, it's an extra step, I've lived without it, or weather would make me worried it might get broken.
That being said, you seem to have a knack for clever ideas and a connection to the community. You are assessing the observations you see happening in real life. That's a cut above the others in my book.
Aside from dumb little devices that I think of in passing, like earrings that are tiny speakers so I don't have to block my ears, I don't have a list of things that I need devices for. Generally it's more like transit and sidewalks and affordable housing.
That being said, you seem to have a knack for clever ideas and a connection to the community. You are assessing the observations you see happening in real life. That's a cut above the others in my book.
Aside from dumb little devices that I think of in passing, like earrings that are tiny speakers so I don't have to block my ears, I don't have a list of things that I need devices for. Generally it's more like transit and sidewalks and affordable housing.
SireSmiggle [OP] 3 points 2y ago
That's very understandable; Most people I have talked to have said that the hassle of learning a new device does not justify the usefulness. Especially if they already can get around just fine with a cane or no assistive devices. Maybe a sensing hat like the one I am making could be useful to deaf-blind? I have never met a deaf-blind person, however, so I wouldn't know.
As for your idea for earring speakers, you may have heard of it, but there may already be a solution in existence. There are earbuds called bone conduction earbuds that do not go in your ears, but around and behind your ears. They pulse vibrations through your temple and into your cochlea. This works in creating sound because sound waves are just particles vibrating, whether it be air or your temple. So you can hear everything around you while also hearing the earbuds.
As for your idea for earring speakers, you may have heard of it, but there may already be a solution in existence. There are earbuds called bone conduction earbuds that do not go in your ears, but around and behind your ears. They pulse vibrations through your temple and into your cochlea. This works in creating sound because sound waves are just particles vibrating, whether it be air or your temple. So you can hear everything around you while also hearing the earbuds.
LadyAlleta 3 points 2y ago
I've never heard of those headphones!!! Thank you!!!!
KillerLag 3 points 2y ago
One of the major brands is Aftershokz, but there are a few others as well. They work pretty well
Superfreq2 2 points 2y ago
Make it discrete, durable, weather resistant, at least somewhat customizable, repairable, and not absurdly expensive, and you probably have a winner on your hands with at least the deafblind community if nothing else.
But marketing to the right people will be a huge part of this, and without your own business, you'll have to partner with someone else.
I really think you still have something to bring to the table here though. You're real world experience is genuinely quite valuable, and it sounds like the smaller projects you were doing before were having more immediate, measurable impacts. You saw an obvious issue, and made steps to ameliorate it. In this case you seem to have a vague idea of what this can help with, but now you run into the solution searching for a problem conundrum, because the application is just too niche and unknown.
I understand and respect the wish to move up to greater things, but honestly their are allot of unknowns there, while what you were doing before was objectively helpful.
What we need right now is iteration rather than invention in most cases, it's not as glamorous or exciting, but it tends to have the greatest positive effect on our lives.
TL:DR: Either find someone who's doing the same thing but is further along and help improve it, or help solve smaller yet cumulative issues like you were before, but don't get out of the field entirely because you have valuable skills to contribute and we still need people like you, just in a different way.
But marketing to the right people will be a huge part of this, and without your own business, you'll have to partner with someone else.
I really think you still have something to bring to the table here though. You're real world experience is genuinely quite valuable, and it sounds like the smaller projects you were doing before were having more immediate, measurable impacts. You saw an obvious issue, and made steps to ameliorate it. In this case you seem to have a vague idea of what this can help with, but now you run into the solution searching for a problem conundrum, because the application is just too niche and unknown.
I understand and respect the wish to move up to greater things, but honestly their are allot of unknowns there, while what you were doing before was objectively helpful.
What we need right now is iteration rather than invention in most cases, it's not as glamorous or exciting, but it tends to have the greatest positive effect on our lives.
TL:DR: Either find someone who's doing the same thing but is further along and help improve it, or help solve smaller yet cumulative issues like you were before, but don't get out of the field entirely because you have valuable skills to contribute and we still need people like you, just in a different way.
SireSmiggle [OP] 1 points 2y ago
Superfreq, reading over your comment hits very close to home. Solution searching for a problem is probably describes what I've got fairly well, I just hadn't realized it until now.
Looking back on it, I've changed this thing so many times in the past 4 years that it doesn't even resemble what it was originally.
My first prototype was essentially a much cheaper clone of one of those text and object recognition devices. I tested an older version of the OrCam MyEye at an event, and I was a little appalled at the $4000 price tag. In theory, OrCam would be a pretty nifty device, but it costs too damn much to justify it's usefulness. When I tested it, it also was just straight up less functional than KNFB, which costs a fraction of the OrCam.
I made a basic copy-cat version of the OrCam that attempted to ameliorate some of the flaws it had. The prototype was most certainly clunkier, heavier, and uglier than the OrCam, but it only cost about $300 to make. I used consumer parts, so if it would ever be manufactured in a factory, I'm sure the price would cut by one-third, or even one half. At $200-300, a device like OrCam would actually be able to justify it's existence when compared to KNFB.
I honestly have no clue how companies like OrCam or Google are making up $4000 price tags for their devices. The physical device itself cannot cost more than $300 to manufacture. Even if they added in costs for marketing and developing software, there's no way it adds up to more than $1000.
After that, I sorta went wild adding random stuff and trying to solve too many issues at once, and I ended up transforming what I have into a navigation device instead.
I could go back to my original idea and iterate upon the OrCam to make something that doesn't require selling your soul to satan to get. However, that may also just be a waste of effort and money as well, because KNFB is really quite nice already. If I do go this route, I could also see a very likely future in which OrCam or Google go after me legally (probably in the form of patent infringement) because I'd be putting a dent in their profits from their overpriced devices...
I'll probably put any development on my hat project on hold until I fully understand if I can solve a real problem with it, or if I'm just hunting for a problem to solve using shiny new tech. In the meantime, I have no intentions of stopping my smaller projects.
EDIT: Damn, I write too much. Forgive me if this is abnormal, I'm a little new to posting on reddit.
Looking back on it, I've changed this thing so many times in the past 4 years that it doesn't even resemble what it was originally.
My first prototype was essentially a much cheaper clone of one of those text and object recognition devices. I tested an older version of the OrCam MyEye at an event, and I was a little appalled at the $4000 price tag. In theory, OrCam would be a pretty nifty device, but it costs too damn much to justify it's usefulness. When I tested it, it also was just straight up less functional than KNFB, which costs a fraction of the OrCam.
I made a basic copy-cat version of the OrCam that attempted to ameliorate some of the flaws it had. The prototype was most certainly clunkier, heavier, and uglier than the OrCam, but it only cost about $300 to make. I used consumer parts, so if it would ever be manufactured in a factory, I'm sure the price would cut by one-third, or even one half. At $200-300, a device like OrCam would actually be able to justify it's existence when compared to KNFB.
I honestly have no clue how companies like OrCam or Google are making up $4000 price tags for their devices. The physical device itself cannot cost more than $300 to manufacture. Even if they added in costs for marketing and developing software, there's no way it adds up to more than $1000.
After that, I sorta went wild adding random stuff and trying to solve too many issues at once, and I ended up transforming what I have into a navigation device instead.
I could go back to my original idea and iterate upon the OrCam to make something that doesn't require selling your soul to satan to get. However, that may also just be a waste of effort and money as well, because KNFB is really quite nice already. If I do go this route, I could also see a very likely future in which OrCam or Google go after me legally (probably in the form of patent infringement) because I'd be putting a dent in their profits from their overpriced devices...
I'll probably put any development on my hat project on hold until I fully understand if I can solve a real problem with it, or if I'm just hunting for a problem to solve using shiny new tech. In the meantime, I have no intentions of stopping my smaller projects.
EDIT: Damn, I write too much. Forgive me if this is abnormal, I'm a little new to posting on reddit.
napoleon88 2 points 2y ago
Honestly this sounds very hackable and quite dangerous if it is hacked. Unless it is completely air gapped, which seems unlikely because firmware updates, you're giving someone the ability to, quite literally, hack your clothes.
Also, in other jurisdictions, there may be issues with someone walking around with a camera that is always filming, though these can be alleviated by ensuring that the device has no storage.
I would be concerned about the privacy implications here. So much wearable tech has been proven to have woeful security.
Also, in other jurisdictions, there may be issues with someone walking around with a camera that is always filming, though these can be alleviated by ensuring that the device has no storage.
I would be concerned about the privacy implications here. So much wearable tech has been proven to have woeful security.
SireSmiggle [OP] 1 points 2y ago
Yeah, safety was a serious concern when making this device. I got a little paranoid when I saw articles about some hackers (good ones, they were hired to do this stuff) who were able to control braking and steering systems in cars by hijacking the bluetooth system on the car's computer.
My primitive solution to this was to make the device a "closed system". Nothing goes in, nothing goes out. Almost all hacking as we know it occurs over wireless connections such as Wifi and bluetooth. If you remove all wireless connections, the only way to hack a device is to physically crack it open and mess with it. That's what I'm going for with my hat.
Keeping it completely airtight, as you said, is a little impossible though. I would like to make this device a one-time-setup deal, so no one has to deal with the cat and mouse game of constantly having to update in order to fix broken shit. However, realistically, I will probably need some way to update firmware, and It will probably have to be wireless.
One solution to maintain security would be to connect the wireless chip to a physical switch on the hat. If the switch is flipped on, the wireless chip has power. If it is off, there is no power running to the chip, and there is no way to power it on other than the switch.
My primitive solution to this was to make the device a "closed system". Nothing goes in, nothing goes out. Almost all hacking as we know it occurs over wireless connections such as Wifi and bluetooth. If you remove all wireless connections, the only way to hack a device is to physically crack it open and mess with it. That's what I'm going for with my hat.
Keeping it completely airtight, as you said, is a little impossible though. I would like to make this device a one-time-setup deal, so no one has to deal with the cat and mouse game of constantly having to update in order to fix broken shit. However, realistically, I will probably need some way to update firmware, and It will probably have to be wireless.
One solution to maintain security would be to connect the wireless chip to a physical switch on the hat. If the switch is flipped on, the wireless chip has power. If it is off, there is no power running to the chip, and there is no way to power it on other than the switch.
BenandGracie 2 points 2y ago
I already do echolocation without any sort of device, so it wouldn't be any use to me. It might work for someone else though.
I don't want to discourage you from creating things because you might just hit upon the next big thing, but the hard part is finding out whether something is useful or not. You might want to find a blind person to test your device to see if it is something that could be useful.
I am not sure if any of this helps, but it is the best I can offer.
I don't want to discourage you from creating things because you might just hit upon the next big thing, but the hard part is finding out whether something is useful or not. You might want to find a blind person to test your device to see if it is something that could be useful.
I am not sure if any of this helps, but it is the best I can offer.
SireSmiggle [OP] 1 points 2y ago
Thank you; I probably titled my post a little wrong.
I have no intentions of quitting from trying to develop devices, I am just unsure if my current sensing hat project is worth time and money. No sense in making something that doesn't actually help the people it is targeting.
Also, knowing how to do echolocation is actually really cool. I've seen videos and articles on people actually doing it, but I've never personally encountered a person who can do it.
I have no intentions of quitting from trying to develop devices, I am just unsure if my current sensing hat project is worth time and money. No sense in making something that doesn't actually help the people it is targeting.
Also, knowing how to do echolocation is actually really cool. I've seen videos and articles on people actually doing it, but I've never personally encountered a person who can do it.
BenandGracie 3 points 2y ago
What ever you decide to do with this device going forward, I like the fact you are asking that question. We see a lot of people around here who have a big idea, and they always seemed shocked when someone points out flaws.
On the echo location thing, I taught myself how to do it when I was two or three years old. This was back in 90 or 91,and my mobility instructor had no idea what it was when I started school. My mom had to explain it to her. I like to think I was doing echo location before it was cool.
On the echo location thing, I taught myself how to do it when I was two or three years old. This was back in 90 or 91,and my mobility instructor had no idea what it was when I started school. My mom had to explain it to her. I like to think I was doing echo location before it was cool.
SireSmiggle [OP] 1 points 2y ago
Wow
At 2 or 3 years old children really can adapt to almost anything. The closest thing I can compare that too in my experience would be swimming I guess. If you toss a kid in the water, they learn to swim pretty quick.
At 2 or 3 years old children really can adapt to almost anything. The closest thing I can compare that too in my experience would be swimming I guess. If you toss a kid in the water, they learn to swim pretty quick.
soundwarrior20 1 points 2y ago
This sounds absolutely brilliant with text-to-speech feedback options of what’s around you this could actually be very useful if you could build in some kind of AI component to guide you as you walking that would also be incredibly useful
SireSmiggle [OP] 1 points 2y ago
So... an explanation of what my little hat is might be useful.
I am using an array of 8 small cameras around a simple black hat with a round brim to gather info. I am using some computer vision algorithms to measure the distance of every pixel on the cameras. I can do this via an effect called stereoscopic depth perception. Having two cameras facing the same direction but a few centimeters apart will allow you to measure distance by comparing the slightly shifted images. Human eyes do exactly this to perceive depth.
This allows the hat to get info about objects in a full 360 degree radius around the wearer, including drop offs and stairs. I am also attempting to add some object and text recognition.
I've managed to make the hat fairly shock resistant (I've smacked it around with a bat a bit), and water-proofing is in the works. I have also tried my hardest to make the hat look sleek and like a normal hat. I wouldn't be caught dead in public wearing some of my earlier prototypes.
My two largest constraints are cost and whether there is a need for something like this.
1. The hat is about $900 just to manufacture, which is stupidly expensive. Hell, I wouldn't pay for some device that damn expensive if I could get around with a cane alright. The blind and low vision people who have tested the device for me mostly agree. Is there any reason anyone here be able to justify using this device instead or in conjunction with a cane?
2. Learning the feedback of a device like mine (vibrations, haptics, sounds) could probably be rough. There's also no "tactile" feedback like a cane.
Any thoughts?
I am using an array of 8 small cameras around a simple black hat with a round brim to gather info. I am using some computer vision algorithms to measure the distance of every pixel on the cameras. I can do this via an effect called stereoscopic depth perception. Having two cameras facing the same direction but a few centimeters apart will allow you to measure distance by comparing the slightly shifted images. Human eyes do exactly this to perceive depth.
This allows the hat to get info about objects in a full 360 degree radius around the wearer, including drop offs and stairs. I am also attempting to add some object and text recognition.
I've managed to make the hat fairly shock resistant (I've smacked it around with a bat a bit), and water-proofing is in the works. I have also tried my hardest to make the hat look sleek and like a normal hat. I wouldn't be caught dead in public wearing some of my earlier prototypes.
My two largest constraints are cost and whether there is a need for something like this.
1. The hat is about $900 just to manufacture, which is stupidly expensive. Hell, I wouldn't pay for some device that damn expensive if I could get around with a cane alright. The blind and low vision people who have tested the device for me mostly agree. Is there any reason anyone here be able to justify using this device instead or in conjunction with a cane?
2. Learning the feedback of a device like mine (vibrations, haptics, sounds) could probably be rough. There's also no "tactile" feedback like a cane.
Any thoughts?
fireborn1472 3 points 2y ago
I would use this. I have no intention of getting a guide dog and a cane takes up a hand that I could otherwise use for something else. It also screems I'm blind to anyone who knows what a cane looks like, so that's another contributing factor. If all the blind people you spoke to said this would not be useful, they are either incredibly lazy and don't wish to learn a new device, or have never needed to carry a bag of shopping and unlock a door :)
SireSmiggle [OP] 1 points 2y ago
Awesome!
I also would not consider the people I spoke to at my local sight center lazy or anything. Most of them simply had no need for a device like mine as they had a significant chunk of their sight remaining, enough to navigate around well.
They also brought up the rational of "It is very hard to make a cane malfunction." It's essentially a fancy stick.
Sticks don't overheat when you are in Quito, Equador and the temperature is constantly over 110 degrees Fahrenheit. Nor do they get in the way if it's -10 degrees and your ears are freezing off. Sticks also don't get uncomfortable when they are wet. A soggy hat on the other hand does.
I've been finding ways to fix these many edge cases, but I just don't think that making my electronic device as versatile as a cane would ever be possible.
I've understood that a tradeoff must be made. I just need to see if that tradeoff is worth it to people.
I also would not consider the people I spoke to at my local sight center lazy or anything. Most of them simply had no need for a device like mine as they had a significant chunk of their sight remaining, enough to navigate around well.
They also brought up the rational of "It is very hard to make a cane malfunction." It's essentially a fancy stick.
Sticks don't overheat when you are in Quito, Equador and the temperature is constantly over 110 degrees Fahrenheit. Nor do they get in the way if it's -10 degrees and your ears are freezing off. Sticks also don't get uncomfortable when they are wet. A soggy hat on the other hand does.
I've been finding ways to fix these many edge cases, but I just don't think that making my electronic device as versatile as a cane would ever be possible.
I've understood that a tradeoff must be made. I just need to see if that tradeoff is worth it to people.
CloudsOfMagellan 1 points 2y ago
Devices like these also aren't likely to pick up on gutters or small steps / dropoffs and will miss bumps and holes in paths as well as short posts and obstacles.
A device like this would almost certainly be used along side a cane, not instead of it
One device I'd like would be some sort of glasses like the Bose frames to integrate with soundscape but then with haptics or something built in so it can act like a miniguide / one of those wrist things
Ocr and image recognition could then be added in future versions
A device like this would almost certainly be used along side a cane, not instead of it
One device I'd like would be some sort of glasses like the Bose frames to integrate with soundscape but then with haptics or something built in so it can act like a miniguide / one of those wrist things
Ocr and image recognition could then be added in future versions
retrolental_morose 2 points 2y ago
I’d pay $1,000 USD for:
• A hat which also had speakers/headphones; I wear a Bluetooth beany in the winter and love it.
• A hat which could learn entrances, landmarks etc and alert me to them. I can’t count how often a bus stops in a slightly different place and I’m then needing to start off from a different point. This could only be a few metres but it has a big impact. Of course there'd need to be a learning process first.
• ORCams facial recognition is clever but impractical, especially when it shouts at you. Being able to set my hat to tap me on the head if it sees someone I’ve told it to look for via an app would make meeting others in public so much simpler. Again, I'd expect to take pictures in the first instance, a bit like Seeing AI's person detection.
• Being able to track an object would be good. I’ll never forget the feeling of walking into a sports stadium with my toddler on my arm, her letting go and me not knowing if she was 2 inches or 2 miles away. How precise could you give distance and direction if I put a tag on her bag, just as a thought experiment?
• being able to toggle detection of overhead obsticles, stairs etc would need mor thinking about. But as a rule, I have a cane for stairs and common sense not to walk into things at chest height.
• A hat which also had speakers/headphones; I wear a Bluetooth beany in the winter and love it.
• A hat which could learn entrances, landmarks etc and alert me to them. I can’t count how often a bus stops in a slightly different place and I’m then needing to start off from a different point. This could only be a few metres but it has a big impact. Of course there'd need to be a learning process first.
• ORCams facial recognition is clever but impractical, especially when it shouts at you. Being able to set my hat to tap me on the head if it sees someone I’ve told it to look for via an app would make meeting others in public so much simpler. Again, I'd expect to take pictures in the first instance, a bit like Seeing AI's person detection.
• Being able to track an object would be good. I’ll never forget the feeling of walking into a sports stadium with my toddler on my arm, her letting go and me not knowing if she was 2 inches or 2 miles away. How precise could you give distance and direction if I put a tag on her bag, just as a thought experiment?
• being able to toggle detection of overhead obsticles, stairs etc would need mor thinking about. But as a rule, I have a cane for stairs and common sense not to walk into things at chest height.
SireSmiggle [OP] 1 points 2y ago
Thanks for the advice.
First off, $1000 is in the realm of too expensive for my tastes. The $900 prototype I currently have built has cameras all around the head, but a lot of people have told me that is fairly unnecessary. I can remove those cameras and get a lower powered computer to cut costs greatly. Based off what you are describing, I could probably make a device that costs about $500 or less. Hopefully less.
Most of what you said is very possible to integrate. The only two things I could see causing serious problems for me would be landmark detection and object tracking.
1. Detecting things like doors or signs are not a problem at all. However detecting new and unknown landmarks can pose some difficulty. Computers learn to recognize objects by looking at hundreds of thousands of images of said objects, and learning what features separate them from other objects. Saving the features of a new landmark would not be hard, but the issue comes when you approach the landmark from a different direction, or approach at dusk where the lighting is different, or a person is standing in front of the landmark, obscuring it. I am not saying that saving new landmarks is impossible though. I will do some research on it and see if I can make a new object recognition model.
2. Object tracking exists and works fairly well. The issue comes into play when the camera is moving around. The camera will most certainly be moving around because it is on a person's head. The moment the camera loses sight of whatever object you are tracking, it doesn't know what to do. This is another thing that I will have to research. Maybe I could save information about how large the object your tracking is, what color it is , and the general shape of the object and have the camera look for those features?
Anyways, this is really useful info to have. Thanks you.
First off, $1000 is in the realm of too expensive for my tastes. The $900 prototype I currently have built has cameras all around the head, but a lot of people have told me that is fairly unnecessary. I can remove those cameras and get a lower powered computer to cut costs greatly. Based off what you are describing, I could probably make a device that costs about $500 or less. Hopefully less.
Most of what you said is very possible to integrate. The only two things I could see causing serious problems for me would be landmark detection and object tracking.
1. Detecting things like doors or signs are not a problem at all. However detecting new and unknown landmarks can pose some difficulty. Computers learn to recognize objects by looking at hundreds of thousands of images of said objects, and learning what features separate them from other objects. Saving the features of a new landmark would not be hard, but the issue comes when you approach the landmark from a different direction, or approach at dusk where the lighting is different, or a person is standing in front of the landmark, obscuring it. I am not saying that saving new landmarks is impossible though. I will do some research on it and see if I can make a new object recognition model.
2. Object tracking exists and works fairly well. The issue comes into play when the camera is moving around. The camera will most certainly be moving around because it is on a person's head. The moment the camera loses sight of whatever object you are tracking, it doesn't know what to do. This is another thing that I will have to research. Maybe I could save information about how large the object your tracking is, what color it is , and the general shape of the object and have the camera look for those features?
Anyways, this is really useful info to have. Thanks you.
sanzaru42 2 points 2y ago
I'd be interested. I'm hard of hearing so can't do the echolocation thing. It'd have to be discreet (or have the possibility of discreet designs down the line) as I have NO desire to broadcast my blindness. Because of the hearing loss I wouldn't want any audio feedback. Also, I don't need 360 radius, 180 would do so maybe you could cut costs there?
SireSmiggle [OP] 1 points 2y ago
One of the main problems I've been trying to address is making the hat seem like a normal bowler's cap as much as possible. I can conceal the cameras quite well in the brim of the hat.
However, unfortunately, I have to keep the processing board and battery in a small bag separate from the hat. It would be a little too bulky and heavy for the head to support alone otherwise. It is a small bag and the cord to said bag can be concealed, but it still is there. It is small enough to be worn under clothes, but still, that's a discomfort that someone shouldn't have to put up with. I am still trying to find ways to conceal it better.
As for the camera radius, I was internally debating removing the back cameras, but the cameras themselves don't significantly affect cost. It is the processing board, batteries, and other electronics (like Inertial Measurment units, tiny haptic motors, etc) that make it expensive. This is still in dev though, so I am always open to ideas to make this not cost an arm and a leg.
EDIT: Completely haptic feedback or nearly completely haptic feedback is very possible. To be honest, it might even be better than having an annoying voice or beep in your ears.
However, unfortunately, I have to keep the processing board and battery in a small bag separate from the hat. It would be a little too bulky and heavy for the head to support alone otherwise. It is a small bag and the cord to said bag can be concealed, but it still is there. It is small enough to be worn under clothes, but still, that's a discomfort that someone shouldn't have to put up with. I am still trying to find ways to conceal it better.
As for the camera radius, I was internally debating removing the back cameras, but the cameras themselves don't significantly affect cost. It is the processing board, batteries, and other electronics (like Inertial Measurment units, tiny haptic motors, etc) that make it expensive. This is still in dev though, so I am always open to ideas to make this not cost an arm and a leg.
EDIT: Completely haptic feedback or nearly completely haptic feedback is very possible. To be honest, it might even be better than having an annoying voice or beep in your ears.
This nonprofit website is run by volunteers.
Please contribute if you can. Thank you!
Our mission is to provide everyone with access to large-
scale community websites for the good of humanity.
Without ads, without tracking, without greed.
scale community websites for the good of humanity.
Without ads, without tracking, without greed.
©2023 HumbleCat Inc • HumbleCat is a 501(c)3 nonprofit based in Michigan, USA.