The popular image processing learning set is now available for everyone to download the "introductory book" for free, but only during the campaign period. The "learning software (IP kit 3)" will be raffled off to those who have downloaded the "introductory book," and the raffle has now concluded, so we will announce the winners. (Previous winners are excluded from this raffle.) This software is available for license purchase through our direct sales or on the Vector website. Direct sales are only available as CD-ROM. If you are interested, you can try out the software for free as a "trial version," so please download it and give it a try. (The trial version has some functional limitations.) (The announcement of winners will list the company names and initials of the last names in no particular order.) Mr. M from Company T, Mr. M from Company N, Mr. M from Company J, Mr. T from Company R, Mr. N from Company D, Mr. Y from Company S, Mr. T from Company H, Mr. M from Company M, Mr. W from Company S, Mr. M from Company D. In total, 10 individuals will receive the software sent to the company names and sections registered with Ipros. We hope you find it useful for learning and practical applications in image processing. *News is distributed via our company newsletter.

It was a question at the exhibition venue. The question was, "With the improvement in CPU performance, image processing development has also become easier, right?" Since cameras and capture boards were also displayed at the same exhibition venue, I provided an explanation while guiding them through those exhibits. As you can see, thanks to advancements in semiconductors, CPU performance has significantly improved. Similarly, the sensors built into cameras have increased in pixel count and improved frame rates, and the manufactured products that are the subjects of inspection using these technologies have also become higher quality. Therefore, I feel that development has not necessarily become easier. Reflecting on this guidance, I realized that due to the advancements in the aforementioned technologies, the demand for companies like ours, which specialize in development, continues unabated. It is not easy, but it is not a struggle either. I believe that this is what allows our work to continue for a long time. If the horizontal and vertical pixel counts of the sensor double, the processing unit will require four times the speed. Even if the pixel count remains the same, if the frame rate doubles, the processing unit will also require double the speed. It seems that competition and improvement will continue in every field. *News is distributed through our company newsletter.

It may be the first time I have been asked to write a name outside my area of expertise in a column title. A "spectral camera" is a camera that incorporates different types of sensors along a single optical axis using optical technology. For image processing professionals, the fact that these different types of sensors can be obtained simultaneously without positional discrepancies is a valuable source of information input. For example, different wavelengths have different characteristics in terms of transmission and reflection. By performing "composite judgment" on these differing characteristics, it becomes easy to distinguish objects that are otherwise difficult to differentiate. In this case, to create a composite judgment method, various ideas are tested and experimented with through software processing. Once the processing is determined, it is hardened (FPGA internal logic computation) to perform calculations at high speed. However, if there are loop processes involved, the advantages of hardening diminish, so the algorithm is modified to enable implementation through pipeline processing. While this may sound simple in writing, it is often a point of heated debate during actual project development. I have provided an example of how advancements in optical technology increase the demand for image processing technology. *News is distributed through our company newsletter.

"Development" is something that cannot be understood without trying it. How do we carry out tasks that involve this? Here’s a guide. Whether on a daily, weekly, monthly, or multi-month basis, the method remains the same. 1. List the tasks. 2. Prioritize the list. This prioritization is not about importance, but about order. This ordering is crucial. For example, if "creating a questionnaire" is on the task list, it involves a counterpart who needs time to respond. Conversely, it becomes a "wait" until the response comes. Tasks like this should be prioritized higher, allowing for other work to be done during the waiting time. As another example, consider "speed research." This is something you investigate yourself, but you predict multiple outcomes. The accuracy of these predictions varies based on experience. The meaning of prediction is to train yourself with statements like, "If it's around **, the next action is ××" or "If it's around ○○, the next action is △△." This practice should be continuously applied and revised for the entire project. I believe this helps in completing the "development" tasks within the planned timeframe. *News is distributed through our company newsletter.

We will be exhibiting at the "Techno Horizon Group Solution Fair 2021 in Nagoya," which will be held at the Nagoya City Small and Medium Enterprise Promotion Hall Large Exhibition Hall from September 29 to 30. Since this is a "private fair showcasing the group's cutting-edge technologies," there will not be a general public attendance. Our company will be exhibiting in the "Machine Vision Solutions" area. We will also be in charge of the "Image Processing Consultation Corner," so please feel free to bring your concerns to us. *News is distributed through our company's newsletter.

The beginning of this was the creation of a textbook for the "Image Processing Seminar" hosted by a seminar company. The seminar takes a full day and covers various specialized areas of knowledge that differ from the development flow, including "fields of application for image processing," "examples of processing," "standard processing," "considering algorithms," "types of implementation methods," "software processing," "hardware processing," "FPGA selection," and "FPGA circuits." We believe that it is difficult to convey this content through materials alone, so we extracted only the first half and are distributing it as an "Introduction to Image Processing" for a limited time during campaigns by Ipros. This announcement is regarding the "update" of this material. Last year, we provided a table for selecting FPGAs with an awareness of pin compatibility, but it was difficult to convey this information through written documents alone, and it was only included in the seminar materials. This time, we have created a table reflecting the "classification of image processing technologies" from three perspectives, which will also be included in the "Introduction to Image Processing" (as an addition). Please look forward to the seminar and the giveaway campaign. *News is distributed through our company newsletter.

Our company conducts image processing seminars twice a year (currently suspended due to the COVID-19 pandemic). Specifically, I serve as a lecturer for seminars organized by the Japan Techno Center. The content begins with an overview of the "fields of application for image processing," but primarily focuses on practical exercises in algorithm development, where participants work on solving problems using established combinations of techniques in the first half. The second half involves exploring implementation methods for the algorithms developed, culminating in the FPGA implementation of the processing. The aim is to provide participants with insights into adjacent fields of their expertise or to help those who are outsourced understand the practical work being done externally. This material is updated annually due to the rapid advancement of technology. The seminar in spring was canceled due to the pandemic, but the materials have been updated. Additionally, the first half of the seminar is offered as a giveaway twice a year through the "Ipros Campaign." The content will also be updated for the next campaign, so please look forward to it. *News is distributed through our company newsletter.

There was a business discussion that started with a customer requesting "consultation on image processing." The content of the image processing matched our company's area of expertise, and the conversation was progressing smoothly. When the customer mentioned wanting to "use the processing in a browser application," considering the effort the customer would have to put into separating peripheral processing, I suggested that our company could take on that responsibility. This was because we had previously handled smartphone app development for customers, especially for internal demos. As we delved deeper into the desired outcomes, both our side and the customer's began to recognize that the "network component" was even more important than the image processing itself. While we take pride in the fact that all our employees are "image processing engineers," we also realized that we only have engineers with experience in "networking" and "smartphone applications" to a limited extent. I conveyed this to the customer and decided to decline that theme. In our area of expertise, just by looking at the problematic image data, we can roughly speculate on the causes. However, in areas outside our expertise, our ability to respond effectively when issues arise diminishes, and I thought that would cause inconvenience to the customer. *News is distributed through our company's newsletter.

Due to the period of priority measures to prevent the spread of infection, the number of exhibitors and visitors was significantly lower than in previous years. However, we were able to exchange business cards with many people who had specific issues they needed help with. I was very worried on the morning of the first day, but I felt that the number of visitors increased on the second and third days. Thanks to the organizers providing many "meeting spaces" where we could move to discuss the overview at our booth, we were able to hold meetings in those spaces, which was beneficial. Of course, since we were discussing matters in such a situation, it is not something that can be easily resolved. Being able to discuss specific approaches, such as "first starting with shooting experiments using this optical system, and then moving on to either this processing or another processing in the next phase," was a significant achievement. As a result of discussing this with one company on the first day, they brought several engineers with them on the final day. The next major exhibition will be the "International Image Equipment Exhibition" at the end of the year. I can only hope that the COVID situation improves and that it becomes a place for more people to "think together." *News is distributed through our company newsletter.

We sometimes receive inquiries from customers saying, "I know what I want to do, but I don't know how to explain it. I've never written a specification document, so please tell me what I should write." We will create the specification document, so please check if it aligns with your requests. Before that, we would like to start by asking you to convey your "wants" to us. Many customers mistakenly believe that when it comes to the actual machine, they will likely use industrial cameras, so shooting video with a smartphone is not helpful for reference. This is a significant misunderstanding. You can effectively communicate "this is what the subject looks like" with smartphone footage. You can also gather a considerable amount of information about your "wants." The specification document will include details like pixel count, processing speed, and accuracy, but the process leading up to that is very important. When we reach the specification document and review it, we will repeatedly ask, "Is this interpretation of your desired processing correct?" We believe that the extent of the back-and-forth communication before starting development on the "specification document" is related to the satisfaction level upon completion. *News is distributed through our company newsletter.

We will be exhibiting at the Image Sensing Exhibition held at Pacifico Yokohama from June 9 to 11, 2021. Our booth number is 22, so please stop by when you visit. This time, our exhibition focuses on "specialized image processing development" and aims to return to the basics with explanations like "a refuge for image processing." There are two exhibits: 1. Image Stitching This is what is commonly known as "panorama composition." It has a special application that is meaningful in achieving this without distorting the images, and we will explain this as a "core technology for individual development" at our booth. To avoid misunderstandings about the incomplete technology, we have combined it with industrial machinery for display, but to bring back the sense of "live action even with freehand," we will also feature some toys after a long time. 2. Simultaneous QR Code Reading While this demo is not new, the highlight this time is the "performance improvement" that is being revealed for the first time. To allow visitors to experience this live, we have prepared a demonstration where items that can be read with the conventional performance and those that cannot be read without the new performance are shown in different colors. Since this is an exhibition, we encourage visitors to interact and genuinely feel the performance. *News is distributed through our company newsletter.

I used a GPU for the first time in a while. The previous case was quite old, but it was about the development of an inspection device using X-rays. In that project, we were in discussions with a company that could "harden" the specified algorithm because we were using a new (higher resolution than before) X-ray camera that was being developed simultaneously, which increased the resolution and processing speed significantly. There were several questions regarding the algorithm the customer wanted to implement. One of them was whether it was acceptable for the neighboring pixels referenced to remain the same despite the increased resolution. Additionally, there was a risk that if we "hard-coded" the algorithm, it would be difficult to make changes later on. Therefore, to meet the required speed, we switched to a proposal to operate multiple GPU-equipped PCs in parallel, which would increase the cost per unit. We successfully achieved the desired speed, but the customer expressed a desire to modify the algorithm. Subsequently, the processing unit of that device was able to achieve processing equivalent to the prototype using GPUs with a single PC equipped with a dedicated board using FPGA, and we successfully cleared the mass production price. There was something I found "amazing" after using it again, which I will share later. *News is distributed through our company's newsletter.

We promote ourselves as a "last resort for image processing," but since it's PR, only the good things are written. Even when we talk about solving difficult problems, we don't have any special abilities. What may seem like "solving difficult problems" is actually just the accumulation of ordinary tasks. 1. Members with a diverse range of different areas of expertise come together. 2. We continue to work steadily on "image processing." 3. We handle a lot of experiences with different purposes for different applications. For example, in point 1, areas of expertise include: 1) being good at objectively grasping things, 2) being an idea person, 3) being good at finding mistakes, 4) being good at forgetting the past, and 5) being good at finding relevant information from an overflow of data. Just having ideas doesn't translate into practical means, and in fact, 90% of the ideas we come up with are discarded. By continuing to focus on each area of expertise and not paying attention to areas of weakness, we steadily build up experiences from different industries in the past, which leads to "solving difficult problems." Although we are a group of ordinary people, being aware that we are a group with the company policy of "one person, one skill" may be the seed for "solving difficult problems." *News is distributed through our company newsletter.

In the development of image processing, the physical sizes of the subjects we target are highly diverse. On the larger side, when it comes to tunnels or roads, the length can be considered virtually infinite (though it is actually finite). Additionally, what we perceive as "large," such as engine parts of large tankers, is impressive, even though they belong to the same category as things we are accustomed to seeing. Perhaps because we are used to seeing car engine-sized parts, it feels like a world from Gulliver's Travels. On the smaller side, atoms and optical elements might be the smallest objects, but thanks to the expertise of physicists and chemists, we can visualize them optically. As image processing professionals, the sizes we directly develop in optical systems might be around "cells," which are in the micron range. For nanoscale sizes, we assist with processing in semiconductors, but the development of the visualized parts relies on the wisdom of various experts. In food inspection, for example, changing wavelengths for foreign object detection has veered off the topic of size, but I feel fortunate to be surrounded by specialists. As long as we can obtain images, we think of various ways to achieve that. This is what makes image processing interesting. *News is distributed through our company's newsletter.

Development referred to the process of stabilizing film during the film camera era, as well as the process of transferring images from film to photographic paper for stabilization. However, in the digital age, camera output has become digital, and it seems to now refer to the process of editing that data using retouching software to make it "look beautiful." Regarding "development" in the digital world, let's consider the function of "making human skin look beautiful." While one might perceive "healthy" skin as beautiful, the definition of "healthy" skin color varies by ethnicity. This presents a challenging example where it is difficult to make "all skin colors look beautiful" through image processing. The display is simple. Showing data without processing it is "display." The true meaning of "display" is to not alter the data. However, human eyes have characteristics such as differing sensitivity to brightness changes based on color. Therefore, when considering a product with the purpose of "television," it becomes necessary to process the image to make it look beautiful before displaying it. There is a distinction between what image processing development engineers think and what product planners consider. I hope I was able to convey the necessity for collaboration between these two perspectives a little bit. *News is distributed through our company's newsletter.