Great Power Technology

Before that, Ye Zhou had carefully read the AES technology documentation provided by the simulator. What he never understood was why the image recognition part in the algorithm module was so large and complicated. At first, he thought that It is to ensure image accuracy, but now it seems that it is not the same thing at all.

This part of the code only exists to solve the problem of neural self-correction.

There is an essential difference between high-precision cameras and the human eye. A very simple particle is that in the civilian field, even the most advanced cameras cannot dynamically adjust the gain of areas with different brightness like retinal ganglion cells. To achieve such a function, you need to use the so-called SME-HDR algorithm.

The SME-HDR algorithm is implemented through a back-end processing system. In this process, a large number of redundant data sets will inevitably appear. These data sets will immediately activate the self-correction function of the human brain to correct it to the correct one. visual signals.

But the problem is that these final versions of the visual signals will also be written back to the camera system, thereby dynamically adjusting the data of the SME-HDR algorithm. The two will continue to interact, and in the end, even if it does not become an infinite loop, it will take up a lot of computing power.

Moreover, the dynamic gain of brightness is not the only problem faced by cameras. Differences in central resolution capabilities, differences in single-photon resolution capabilities, too mapping algorithms, differences in color and color gamut sensitivities... Various problems are added together to make corrections difficult. The difficulty of making mistakes increases exponentially.

Computers can handle this amount of calculations, but what about the human brain?

Therefore, Yang Yijin could already see Yaoyao's discomfort when the experiment was only halfway through - because she was indeed very tired at that time.

Ye Zhou sighed deeply. He had never thought that what limited the development of AVS technology was not the technology itself, but a factor that almost everyone ignored: the camera.

Moreover, the problem is not that the camera is not advanced enough. On the contrary, it is because its functions are too advanced for the human eye.

The resolution is too high, the color distortion is too small, and the shape distortion is almost non-existent...

If you want to express it in plain language, the visual system of a normal person is a third-rate camera paired with a first-rate processor, but under the AVS system, it becomes a super-first-rate camera paired with a second-rate processor, because according to With the current solution, the neural pathways established by the AVS system cannot reach the agility of the original neural pathways of the human body.

With one thing going the other way, it becomes natural for problems to arise.

There are only two ways to solve this problem: either customize a special camera or perform invasive surgery to widen the nerve path.

However, the increased costs and decreased convenience brought about by these two methods are extremely unsuitable for large-scale promotion.

Therefore, the final solution became to use algorithms to compensate for the defects of the equipment.

This obviously violates the principle of "code should be as concise as possible" in computer science, so it actually has a negative impact.

Thinking of this, Ye Zhou shook his head slightly.

It has to be said that the researchers in this timeline are lucky. This luck is not because we directly acquired the AVS technology, but after acquiring the technology, we selected the most typical self-correcting system in the first experiment. The subjects with the strongest error ability.

Such luck allowed the most extreme situation to be tested in one go, but on another timeline, thousands of tests might not be enough to achieve such an effect.

After listening to Ye Zhou explain the cause of the problem, Chen Li's team was also surprised, especially Zhu Xuan. He had already discovered a large number of redundant codes in the algorithm and analyzed its functionality, but he always wanted to I don't know what problem these codes are used to deal with.

Now, all your doubts have answers.

"So, if we want to continue to develop AVS technology, we have to start by building cameras first?"

Hearing Chen Li's question, Ye Zhou nodded slightly and replied:

"Either build a camera or do a craniotomy. The risk of the latter is too high. You can't just open a hole in the back of your head and connect a USB port like in that kind of science fiction movie, right? The risk of infection is no joke. .”

"Okay...but as a result, this experiment will be postponed. It is not difficult to build a camera with functions similar to the human eye. The main difficulty lies in the debugging process. I estimate that this process will not take more than three to five years. Can’t get off.”

Chen Li looked a little disappointed. His eyes seemed to glance at Yaoyao who was sitting quietly aside. Everyone could see what he was thinking.

The cruelest thing in the world is not losing hope and then having it taken away in a short period of time after regaining it.

There is a sentence that says, "I could have endured the darkness, if I had never seen the light." This sentence seems a bit pretentious, but it is a true portrayal of Yaoyao.

She had seen light, literally, but now, she had fallen back into darkness.

Ye Zhou stepped forward and touched Yaoyao's hair. She looked up in confusion, as if she still didn't understand why she could see it just now, but now her eyes turned gray again.

"Uncle, is your computer out of power?"

After hearing her words, the expressions of everyone around them were a little bitter. Only Ye Zhou smiled and replied:

"It's not that there's no power, it's that your new eyes need to rest."

After finishing speaking, he did not look at Yaoyao's reaction, but turned directly to Chen Li and said:

"Don't make the problem so complicated."

"Developing new cameras is something that needs to be done, but this does not mean that the action of helping blind people regain their sight has to be suspended."

"The current problem is very simple. There is a conflict between the computing power of the nervous system and the density of the camera's information instruction set."

"In that case, it would be nice to reduce some of the information set."

“The information set that removes color and depth of field, leaving only shape and light and dark.”

"Isn't it also a huge progress to make blind people completely color blind?"

….

Two hours later, the AVS system after removing the color and depth of field lenses according to Ye Zhou's requirements completed the test, and the test results surprised everyone.

Before this, they generally believed that human beings' judgment of depth of field came from binoculars, that is, dual cameras, but it turns out that the adaptability of the human brain is much stronger than they imagined.

Threading a needle with one eye is difficult, but not impossible.

Yaoyao obviously has amazing talent in this aspect.

Just over an hour after the experiment started, she was already able to walk in the room, go up and down steps, and avoid obstacles with the help of visual signals transmitted by the AVS system.

This seemed to be effortless for her, and even when Chuya helped her remove the interface electrode, she seemed a little unfinished.

"Okay Yaoyao, please sit aside and wait for a while. We have to assemble the equipment so that you can take it away."

Yaoyao nodded obediently and groped and squatted on the ground. Yang Yijin walked over and picked her up and placed her on the sofa outside the laboratory.

The assembly work was handled by several junior researchers under Chen Li. He did not need to do it himself, so he sat with Ye Zhou and others.

"What you started doing was a mobile solution?"

Ye Zhou looked curiously at the AVS equipment that was gradually assembled into a helmet style and asked.

"It must be a mobile solution. Otherwise, if the experiment is successful and the subject cannot take it with her, that would be nonsense... Moreover, the process of information collection is not over yet, so let her take it with her to a complex environment. Further optimization of the algorithm will actually be helpful for subsequent equipment improvements.”

Ye Zhou nodded and did not ask any more questions. Instead, he turned to Yaoyao aside and asked:

"Yao Yao, do you feel accustomed to your new eyes?"

Yaoyao on the sofa frowned and thought for a moment before replying:

"Colorless."

"Yeah, you can only have color if you can."

"But I want to see color."

"But there's nothing you can do now. If you want to see colors, your head won't be able to bear it."

"I can stand it!"

Ye Zhou burst out laughing. After all, she was still a child at heart, and when she encountered what she really wanted, she would still act like a spoiled brat like her peers.

Thinking of this, Ye Zhou did not continue to entangle with her topic. He just deliberately paused for a moment and continued to ask:

"Now that you can see, what do you want to do?"

"Go to kindergarten."

Hearing this answer, everyone on the sofa laughed.

After laughing, I felt a little sad again.

The accident at home actually happened when she had just finished kindergarten, so her impact on "schooling" probably only stayed at that time.

She can learn a lot from her mobile phone, but she cannot learn the concepts of elementary school and junior high school after kindergarten.

"You are almost seven years old. You can't go to kindergarten anymore. You should go to elementary school."

"Then go to elementary school."

"What about after you finish elementary school? What about when you grow up? What do you want to do?"

This is a classic question. It seems that when all adults face children who have just started to go to school, they can't help but ask "What will they do when they grow up?", just like when facing children who have just learned to talk and walk, they always ask It's like asking him "Where are mom and dad?"

Ye Zhou thought she would give answers such as astronauts and scientists, and even thought nonsensically whether she would say she wanted to be a lumberjack, because yesterday at her home, he saw the animation on her phone. The film is about lumberjacks.

But unexpectedly, her answer was different from what everyone guessed.

"I want to be a firefighter."

Ye Zhou was stunned for a moment and asked:

"Do you know what a firefighter does?"

"I know."

Yaoyao's face showed a serious look like a little adult. After a moment, she lowered her head, slowly intertwined her fingers, and then whispered:

"Firefighters put out fires...fires...are bad things."