Search academic texts

Information × Registration Number 0226U002289, (0123U100012) , R & D reports Title Development of new adaptive information technologies for recognizing objects in the surrounding environment popup.stage_title Розробка нових адаптивних інформаційних технологій для розпізнавання об'єктів навколишнього світу Head Koryashkina Larysa Sergiyivna, к.ф.-м.н. Registration Date 13-02-2026 Organization National Technical University Dnipro Polytechnic popup.description1 Development of new adaptive architectures, blocks of neural networks, as well as learning procedures for effective recognition of objects in the surrounding environment, including on devices with limited computing resources. popup.description2 A set of adaptive information technologies for intelligent recognition, analysis and recommendations in real time has been created. The adaptive PAN+PTA Edge-Seg semantic segmentation technology is used to solve the problem of weed recognition. It is implemented on peripheral computing devices and effectively adapts to devices with different computing capabilities without retraining. The FoodVision-LLM information technology is designed for ingredient segmentation and automated nutritional analysis based on a two-stage approach that combines an adaptive neural network for ingredient segmentation in an image with a large language model (LLM), which converts information about ingredients into nutritional data and has a chat interface. The Spark-RecoSys technology is used to generate personalized recommendations in distributed data processing systems. Product Description popup.authors Kostiantyn S. Khabarlak popup.nrat_date 2026-02-13 Close