Adam Dubis

Title: Adam Dubis: Medical Imaging and Object Detection Innovations.

Introduction. Adam Dubis is an inventor based in London, Great Britain. His patent work addresses neural-network methods for analyzing medical images and improving object detection in digital images. He is associated with UCL Business Plc and has 2 patents.

Latest Patents. “Medical Image Analysis Using Neural Networks” describes systems and methods for identifying tissue layers and their boundaries in medical images. A neural network assigns boundary-surface scores to image pixels to estimate whether they correspond to boundaries between segmented layers. One embodiment uses two connected subnetworks: the first identifies tissue layers, while the second determines the boundaries between them.

“Computer-Implemented Method of Enhancing Object Detection in a Digital Image of Known Underlying Structure, and Corresponding Module, Data Processing Apparatus and Computer Program” describes a method for improving object detection when an image has a known underlying structure. The method uses positive images containing detected objects and negative images without detected objects. A neural network extracts and contrasts feature vectors from paired images, forms an attention vector, and processes it with the positive image’s feature vector to produce an output.

Career Highlights. Dubis’s listed inventions apply machine-learning techniques to two distinct image-analysis challenges: identifying anatomical layers in medical imagery and detecting objects against a known image structure. His work is associated with UCL Business Plc.

Collaborations. Watjana Lilaonitkul and James Willoughby are identified as Dubis’s coworkers.

Conclusion. Adam Dubis’s patent portfolio focuses on neural-network approaches to medical image segmentation and enhanced object detection, reflecting applications of machine learning in image analysis.