VisuelAIclassification, a tool to share your visual skills without coding. Test on a new hematology dataset

Image classification is one of the key fields in machine learning and computer vision. However, building, training, and testing image recognition models often require extensive programming skills and computational resources, making the process challenging for many researchers and practitioners. For example, in the medical domain due to lack of technical expertise 71% of the institutions report barriers to AI implementation. This article presents the software VisuelAIclassification, which includes an intuitive and guided graphical user interface (GUI) built with CustomTkinter. The software is designed to, on almost every computer, to prepare dataset (data augmentation available), train machine learning models, test machine learning models, preprocess images by removing metadata, and sort images with easily shareable annotations. The primary aim is to provide an accessible tool for researchers and practitioners in the field of computer vision. An important hematology dataset is also delivered to demonstrate the software strengths.