I frequently give talks on a variety of subjects (including entrepreneurship, creativity, robotic vision…) It always takes an extra effort to coordinate some images with the material I’m trying to put across via Powerpoint presentations. What is sometimes lacking, though, is an element of extra spontaneity.
Today’s invention is a tool to make talks more attention-grabbing and possibly memorable.
As the presenter speaks, his words are analysed by a computer undertaking rapid speech-to-text analysis. It then performs a realtime image search and displays on the screen one image from, say, the first page of results. With suitable filtering in operation, the scope for embarrassment would be limited and the process might even result in some interesting feedback -if the presenter were to watch the images and comment on them too.
For many talks, it wouldn’t matter if the translation wasn’t that accurate or the recognition rate was low -just as long as images appeared broadly in synch with the main words. This might have the added benefit that speakers would be encouraged to enunciate more clearly (try Googling ‘eh’ and ‘um’).
Finally, this technique also represents a way to add some extra interest to radio broadcasts -it’s a kind of pictorial stream of consciousness rather like the one which visual thinkers naturally adopt.
It’s sometimes quite disturbing to see the amount of food wasted in a large canteen when the trays are returned to the trolleys. Maybe people turn out to be less hungry than they thought they were, but I suspect it’s due to a combination of standardised portion sizes and unappetising fodder.
Today’s invention attempts to address the latter problem.
Trays would be provided, as often happens now, with a space for each component of a meal (ie meat, vegetables, salad, dessert, drink container etc). They would be made in translucent material. Servers would be shown which recess should accommodate which component for each sitting. (The trays themselves might actually be available in big-recess and little-recess variants, in order to limit the first problem mentioned above).
After each meal, trays would be returned to a conveyor belt for cleaning, instead of being piled on trolleys. The conveyor would allow trays to be placed on it in only one orientation (by the use of eg moulded-in lugs). Each tray would then be illuminated from below and viewed by an overhead camera. The amount left of each component could then be measured by standard image processing techniques.
This would allow a very rapid assessment of what parts of a meal were well received and thus guide future menu design in order to lessen waste in a very cost-conscious application. Evidence of a failure to choose eg salad or fruit could then also be used to drive healthy eating education programmes.
When making a braided rope, a process akin to plaiting takes place which can be thought of as three strands at a time being linked together.
Rope and braids exhibit the emergent property that their load carrying capacity is greater than the sums of the capacities of the individual fibres. Braids tend to fail, however, when weaknesses in different fibres, which are randomly distributed along the length of each one, find themselves in alignment. To help reinforce such weaknesses, the fibres are woven regularly and tightly, but this limits flexibility.
In a 1-D cellular automaton (CA), the state of three adjacent cells specifies the state of the cell at the next level down in the pattern (See here). This is not unlike a process of braiding.
We can interpret whether a cell is black or white as meaning that the corresponding fibre in the braid is ‘over’ or ‘under’. (In this way, a conventional plait rule would look like 001-> 1, 100->1, etc…) The amazing thing about such CA systems is that they can generate something very like randomness, using only simple starting conditions and rules. Rule 30, as shown, is one such randomlike rule.
Today’s invention is therefore to use Rule 30, or similar ones, to form highly flexible braided ropes in a deterministic way.
(The sequence of fibres in the diagram starting at the top centre position and running vertically downwards would be something like …oouuouuuooouoouu… (with ‘o’ meaning over and ‘u’ meaning under) ie a continuous but almost random, and therefore highly flexible, weave).
Ants are remarkable creatures in that they exhibit certain cooperative behaviours which ’emerge,’ based on the interactions of some very simple internal rules. Some can, for example, count the numbers of steps they take, in each direction on some irregular route, in order to compute the distance and bearing of the most direct path home.
Certain species of ant, it seems, are able to measure themselves against the size of holes in the path of an army and pick appropriately-sized individuals who can then use their legs to plug the potholes.
This behaviour might actually be exploited commercially.
If you have a natural product, as the sponge fishing industry does, it’s often quite hard to gauge the quality objectively (certainly, doing this by imaging techniques is fraught with problems). Given the prices which individual sponges fetch ( >$30 each) , this is a significant issue.
Today’s invention is a pore size measurement technique. Allow an army of ants of the pore-filling variety to march across the surfaces of the product, leaving those ants with legspans equal to pore sizes in the (sponges’) surfaces. These creatures can then be dislodged and measured (either individually or by a statistical analyisis of post mortem leg length.)
If you’re driving a long way and still feeling fresh, the last thing you need is to have to draw into a service station and buy fuel (together with their wildly overpriced coffee, sandwiches and traditional butterfudge cake).
If aircraft can refuel in-flight without needing to drop into the nearest Moto station then why can’t cars? (btw, who the hell designs Moto signage, Martians? )
Today’s invention is en-route refuelling. When your vehicle’s tank approaches empty, rather than having to stop and lower your average speed, you can simply call up and locate the nearest road tanker vehicle. These would be hammering along the world’s road networks full of eg four-star, diesel or LGP. A distributed computer system would detect which tanker was nearest and direct both vehicles towards a rolling rendez-vous.
On meeting up, you would drive your vehicle until it docked with the rear of the tanker and received an injection of fuel. You might even be able to take on some window-wash or radiator water -or even a cup of coffee (toileting facilities might be made available, but I’ll leave those details to your imagination).
Refilled, your vehicle would undock and carry on its merry way.
I found myself standing recently on a city street where I used to live and I couldn’t believe the level of pollution in the air. It was actually hard to breathe and the whole area stank -mostly the smells of vehicle exhausts (although there was also a whiff of abandoned rubbish, since the council decided to save money by going for bi-weekly collections).
We routinely have cars now with catalytic convertors (and they smell pretty terrible too), so I wondered, why can’t cars smell better?
Today’s invention is to fit activated carbon filters (ie the ‘odour-eaters’ used in trainers) within vehicle exhaust systems. This could seriously reduce the unpleasant smells emanating from the rear ends of cars and trucks. A further improvement could be made if we arranged to inject ‘natural’ scents into the exhaust stream of each vehicle (cut grass, perhaps or baking bread or newly tilled soil…).
This does nothing, of course, about the levels of particles emitted or dangerous gases but it might make urban life more tolerable while we work on solutions to those other issues.
I’m becoming increasingly frustrated by the quantities of dust I’ve been finding on all the optical kit to which I have access. Thinking about how best to remove or control the problem has led me to do some background research.
Then I started to think about whether this problem could be converted into some kind of benefit.
In fact, if you have a pattern of dust particles on a glass screen protecting a sensor in a camera or scanner, then that pattern could be used as a ‘watermark’ to help defend your Copyright in the resulting images (assuming you were the kind of unrelaxed type who cares about that kind of stuff more than taking great pictures). This is today’s invention.
No one wants visible particles in their images, but it might be possible to take a calibration image of a very uniform field and thus establish the distribution of even invisible dust particles. This could then be actively sought, using statistical techniques, in all subsequent images taken.
The calibration image would need to be captured through the lens system and might be achieved by taking an image of the inside of the lens cover by briefly firing the flash, within the camera body, every time it is opened. This would probably require storing coordinates of the biggest 20 particles present and associating these values with the images taken in the current shoot.
If relying on natural dust ingress is an unattractive option, then I’d propose that every sensor array be manufactured with a unique distribution of minute biases (or that natural biases be recorded). These would be invisible when the images were viewed by a person, but could be detected by statistical image analysis techniques if Copyright, or tampering, were later at issue.
One of the biggest commercial questions of the age is
“How the hell can we get those people to stop fast forwarding through our beloved TV advertisements?”
Todays’ invention attempts to deal with the underlying problem. No, I don”t mean the really underlying problem that companies still think it’s a great idea to force their potential customers to watch promotional material when they are trying to be entertained.
The thing about fast forwarding through advertisements is that you really have to pay close attention; otherwise you find yourself skipping into Part 3 after the break. So, imagine shooting a couple of seconds of mind-blowing ad action and then broadcasting it in slo-mo.
First, somene watching the ad in realtime will think “What’s going on here, has my recorder blown a circuit?”…so, now you’ve got their attention. Then, they will do the usual thing of fast forwarding…at which point they will see a short burst of normal promotional action. “Acme sugar water -get’s you up to speed.”
Sneaky? and of course it will only work a few times, but it might just keep a few of those Madison Avenue types in in Dom Perignon for one more season.
Hairstyling is a matter of big emotional significance for surprisingly large numbers of people. Experienced coiffeurs can tell at a glance whether a particular style will suit any given individual, but for the customer it’s difficult to choose appropriate designs.
Sure, you can cut and paste a static haircut image onto your passport photo at a number of online venues, but somehow the effect is never that convincing -or reassuring.
Now that animation experts (eg at Pixar) have managed to solve the research question of how simulated hair moves, today’s invention is a way for people to use that information to plan their new hairstyle.
A webcam image would be taken of the customer’s crown, to indicate the basic layout of their hair growth pattern. This, together with some information about hair type (eg, wiry, greasy, etc) would be used as input to a program which would create a model of their current hairstyle.
Armed with this information, a stylist could simulate the process of cutting the hair into different shapes (recording the process for training purposes, perhaps) and then supply a customer with a range of fully animated options.
(In theory, it would then also be possible to have the haircut undertaken remotely. The customer could place their head inside a electrostatically-charged sphere, with radius larger than the longest hair, which would cause every hair to stand on end and be clipped to the length specified in the animation model. This would only be for intrepid and/or impatient types, though, since I’m informed that watching a hairstylist perform the cut is actually a valued part of the experience).
Walking boots are now available in a huge variety of colours, ‘technical’ materials, weights and prices. It always bugs me though that when I strap on my old Karrimors, the heavily-cleated soles quickly fill with mud and don’t provide me with any more grip than a pair of Italian loafers (not that I’d ever wear these, even if they came in my size).
It’s perhaps not quite that bad, but there’s certainly some kind of ongoing design compromise by manufacturers so as to provide a certain amount of traction, at a reasonable cost but without destroying the surfaces of woodland paths and rock formations all across the globe.
Today’s invention is a new approach to walking boot soles. Instead of fixed cleats, I’m proposing an array of retractable studs (resembling about fifty short golf tees per boot). These would be made of eg tungsten and free to move independently up into the 2cm-deep base of the sole, but each would be mounted on a small spring. This would allow the sole of the boots to conform more closely to the local shape of the ground surface and thus provide much better grip. They might also give a certain extra protection from damaging impacts when clambering over rock surfaces.
Walking on a smoother surface afterwards would force more of the studs upwards and allow any accumulated mud/debris to be sloughed off (without having to wait the usual week for it to form dry cakes and fall away).