Object based video search?

I'm sure this is already done/being done. But what the hell.

Today's video search is based (only?) on the text that the uploader describes a video with. Can video search be smarter by recognizing objects in the video? Give me results based not only on the text that the uploader has included but also using the content in the video.

Going somewhere ? I'll deliver it for you.

Scenario: I am in a movie, I need some of my things to be transported to a particular place in the city. If there could be a service to pick my stuff and deliver it to the concerned person, I would be so relieved.

Day in and day out, I would always have some stuff to be transported to some other place in the city.

I might want to courier some stuff. But they don't come to me. I wish they do.
I might want to post something. I wish somebody pick up the posts from me.

Okay. So the goal is provide service that could transport anything within the city.

Disadvantages: We only have very less scenarios like this. Most of the times, if we might want to take some money or valuables (like Passport :P), we would transport it ourselves.

But if am ready to pay for it, I wish there be some service.

Implementation: We could initially target a famous courier service in the city. All those people who would have to courier something. We say we could go to them and pick up the stuff.
We might want to target a very big post office too.

Experiments for all!

Premise:
Education is changing for good with explosion of technology and mobile phones. Can we help students learn more efficiently and effectively than before. I do not know of other things but, I think it is possible in Physics.

Problem Definition:
Many a students and schools across the globe do not have any experimental equipment to learn science the way it is to be learnt. Even in schools where there are good labs, students do not approach labs with curiosity, probably because of their mono-tonicity.

Solution/Thoughts:
A single platform for science based experiments as a game. E.g: We can have a few tools like pulleys, battery, motor, copper wire, magnets, ropes etc., and ask students to build a solution for a problem like moving a 20kg stone from one place to another. We could ask them to build/engineer things using these basic tools and achieve an objective like design an harpoon to scale enemy walls in a war-like game. We can have multiple stages with each stage based on a concept with a set of tools.

Existing games:

There are a few games existing on somewhat similar lines. But none which is a generic platform for science experiments:

Angry Birds: A very successful(>100Million downloads) game on iPhone/Android with which students can learn about projectiles, even though it was not designed with the purpose of teaching Physics.
http://fantasticcontraption.com/: This is a more exhaustive Physics simulation game which was launched in 2008. In 6 months more than 3 million people played it. It certainly teaches some engineering/building principles, but nothing in a standard education set-up.

Technologies needed:
1> Android/iPhone Programming
2> Physics engine
3> Designing a gaming framework from scratch for science.

I am not looking at this as a money making idea. But something that is interesting and satisfying to be part of. Any one ready to be part of it?

A new way of selling Art

Hi all,

I hope every one know about firefox personas. We can design an addon, which shows the paintings or photographs as a browser wallpaper. So If people like the browser wallpapers(I mean paintings or photographs), they can buy it online, and user can also rank the paintings or photographs. I believe this would be a good platform for artists, so they can reach millions of people using our addon. We can enhance it to other domains, like fashion.

Thanks,
sathi.

Picture-Perfect

Problem: A problem I see around very often is people's need to get "great-looking" photos. It has become a fixation for most of us especially after the Facebook-profile-pic  craze. The question is "Can photos taken by amateurs be made great looking with little effort and no specific skill-set?". Imagine doing this editing in an iPhone/Android.

Current-Approaches: Many photo sharing apps on mobile phones(e.g: instagram) have a simple yet very useful feature called "filters" which apply predefined effects on images. They actually do a  good job and people simply love them.

Proposed Solution: Let us understand what is meant by a good photograph:

1> A very nice focus on 'main' object and blurring effect on the background=> Lens-Blur

2> apt-cropping of the image (or) neat framing.

3> Color-effects like in filters above.

4> Removal of unnecessary artefacts or noise. 

Some of these things can be done in a semi-supervised approach with some input from user and some of it done automatically.


Business-proposition:   You can give away a few effects for free and have a few paid features. With million of camera-enabled phones and with photos being a primary usage for them, you have a huge pool of buyers.

Global Hotkeys

Ever wondered how many key combinations our keyboards (assume a 101 key keyboard) have and what all we can do with those?

The answer to the first part of the question - I leave it to you guys to answer.
My answer to the second part - A LOT!

For the benefit of those who don't know what global hotkeys are, these are key combinations assigned to perform a task in an application. Just like Ctrl + Alt + Delete. Another example, I'm currently using the browser and I want to switch to the next track in my music player. A simple Ctrl + Alt + Right Arrow does it for me. It avoids the arduous task of going to the music player and switching to the next song and returning to the browser. Got the hang of it? Good.

If there's a way to assign custom keys to various tasks in various applications, wouldn't that be cool? Now add mouse input to this. Even cooler. Add touch to it. Add speech to it. Add any other input mechanism you can think of to it. Too complicated now? That's ok. You'll get used to it. You'll be amazingly productive on the computer thereafter.

AFAIK, there doesn't exist a tool through which you can assign shortcuts for tasks from various applications even for the keyboard. Such a tool would a super cool tool in my eyes.

PS: Quite an off-topic post probably? Just wanted to let people about global hotkeys if they haven't heard of them.

Server-Server #waiter

• Pain points
• To customer: You go to a restaurant and order something the first time. Then afterwards, for ordering the next item (which will typically be extra naans or some new thing). But calling the waiter: "hey bhaiyya"/"waiter" or some other thing is a problem. Even bigger problem is when you don't remember who your table's waiter is - and you want some spoon/onion or something.
• To customer: You don't know how each item looks in the menu. Typically you ask the waiter about specific items: "Bagara baigan mein gravy hota hai kya?" kind of questions. The pain here is that the number of questions you can ask the waiter is limited (you can imagine why) and you cannot see how every item looks like.
• Also, if there are some suggestions about what is special/quickly available/generally preferred item in a particular restaurant, it will be helpful to the customer and the restaurant owner will be happy to provide this data for the users.
• Idea
• What if all the menu cards/restaurant tables have QR code or some marker? When customers arrive at the restaurant table, they can point their smartphone cameras to the marker, which fires the menu up. The menu will be a standard card type menu, where items are clickable. 
• On clicking a particular menu item, you can see the image of how it looks like (solves the gravy problem), quantity information (veg biryani in different restaurants serves different number of people), etc.
• When you want to order another item from the middle of your meal, you order it from the app itself (you give the quantity and item), you know the running price (track to see if it is in your budget), you get suggestions about what is popular in this restaurant, etc. Basically, computer server becomes the hotel server here...
• What more? The order you give fires up a timer, if you do not get service within a time, you can rate the restaurant directly from there (think improved service).
• Once we get all the rating/orders/popular-items-in-a-restaurant, etc. we can provide analytics about it to the restaurant owner about what items are going well, what needs to be improved, how is the service at each table, etc.
• Since we know all the analytics, we can provide discounts to repeated customers (configurable by the restaurant owner).
• Technology challenges/description
• The augmentation part is easy.
• Analyzing the data and giving useful feedback to restaurant owners is a big data/text mining challenge.
• Marketing this app to users and restaurants.
• Competitors
• Yelp/OpenTable (not direct competition though), FourSquare (in the discounts category).
• Market 
• All restaurants which expect customers who have smartphones with cameras.
• Business model 
• The app can be free both to restaurant owners and customers but the data of which item is being ordered by whom, and how many times (data analytics/market research, basically) can be sold to other companies or even restaurant owners themselves.