Surgery Godfather Chapter 1393 - 1393 1052 The Real Tycoon

Chapter 1393: Chapter 1052: The Real Tycoon Chapter 1393: Chapter 1052: The Real Tycoon Graduate students from the Surgical Research Institute and researchers from the Digital dical Laboratory, everyone helped review the images. Unfortunately, after looking through thousands of images, only one image showed structures of a virus within the cells, and no traces of tumor cell apoptosis were observed. In other images, whether they were of tumor cells or normal cells, no structures of the K virus were found.

These are all within Yang Ping’s expectations because experints cannot always go smoothly; there are bound to be various problems during the process. The success of an experint is a continual effort of solving problems and adjusting strategies until final success. Many experints, despite imnse efforts and countless tasks, still end in failure which is common.

Moreover, now that one image has shown the structure of the K virus within a tumor cell, this is already a huge success, indicating that the K virus can successfully reach and infect tumor cells.

There are several possibilities for the current situation; first, most viruses might be eliminated by the immune system during their transport in the blood to the tumor. Second, a lot of viruses could reach the tumor, but the success rate of infecting tumor cells is extrely low. Third, the viruses that infect tumor cells may initiate the apoptosis program, but the cells that start the apoptosis program were not captured. Fourth, it’s possible the viruses can’t initiate cell apoptosis.

No matter how many possibilities, one thing is certain: the K virus can infect tumor cells, it’s just a matter of how strong or likely it is, and this image right now is the strongest evidence.

Tens of thousands of images were captured by the cara, and reviewing them manually is too inefficient and also unhealthy for the eyes.

As the saying goes, “Sharpening your axe will not delay your job of cutting wood,” Yang Ping decided to use computer assistance, and he suggested that Director He’s team develop an artificial intelligence model to help differentiate these images.

In fact, for Director He’s team, developing such a model is not difficult because they already have an AI model for image recognition, mainly used for pathology slices and image photo reading, and this model is very mature in recognizing pathology slices and images, already surpassing the majority of doctors.

Doctors might find it exhausting to view hundreds of pathology slices a day, but AI could possibly read tens or hundreds of thousands of images in just a few seconds. Such efficiency is already being utilized in dicine, and many hospitals can fully automate the viewing of pathology slices in the future, without needing pathologists to review them. If the production of pathology slices could also be automated, pathologists would only need to handle pathology specins simply.

Any artificial intelligence needs a large amount of data to feed, otherwise, the AI is just an “idiot.” In this experint, there are no precedents to reference for recognizing the images of cells from electron microscopes, so there is no data to feed the AI.

Although Professor He’s team using Digital Human Technology has developed an imaging AI model specifically for intelligent film reading, such as X-ray, MRI, and CT scans, this model has been fed a large amount of data and has evolved nicely.

However, for images of tumor cells under an electron microscope, these are entirely new images, lacking extensive data to feed, so the artificial model cannot evolve and is not very useful.

Professor He could only develop an image analysis engine based on Yang Ping’s summarized experiences, which doesn’t need large data to feed; it could analyze images. If the analyzed structure matches that of tumor cells without any additional structures, the image is marked as uninfected.

If extra structures are found inside the cell, they are imdiately extracted into the next analysis program. If the analysis matches the shape and structure of a virus partially or fully, it is marked as infected.

If uncertain, the images are marked as unclear and will be left for manual analysis. Afterwards, based on the results of manual analysis, experience is summarized and applied to refine the model, thus making its engine more advanced and increasing its recognition capability.

This is sowhat similar to how antivirus software analyzes computer viruses; the difference is that antivirus software analyzes the code of computer viruses, whereas this software analyzes electron microscope images, but the underlying principles are similar.

In short, Yang Ping now needs to closely cooperate with Professor He’s team to develop various specialized models to assist with the experints. When the full-scale AI model is ready, it will use the large model to aid research, greatly enhancing research efficiency.

Developing a model by Professor He is not a task that can be done in a day or two, although he already has a base model acquired that his team has managed to digest and absorb. Modifying these models also takes ti.

When Ruixing dical acquired the model, it included models specialized in lab biological image recognition; Professor He’s team could improve this model, thereby saving a lot of ti.

Actually, acquisition is not a bad thing; if post-acquisition absorption and digestion of technology are achievable, it would greatly shorten the research and developnt ti, maintaining synchrony with international standards.

Yang Ping decided to temporarily stop viewing the images until Professor He’s image recognition engine is developed and then continue with the remaining images. He wanted to focus his main energy on modifying the virus’s structure.