Adaptify: A Refined Adaptation Scheme for Frame Classification in Atrophic Gastritis Videos
This work addresses the challenge of model reliability in medical video analysis for atrophic gastritis detection, which is an incremental improvement in a domain-specific application.
The paper tackles the problem of inconsistent reliability in machine learning models for atrophic gastritis detection in videos by proposing Adaptify, an adaptation scheme that integrates knowledge from an auxiliary model, resulting in notable improvements in output stability and consistency.
Atrophic gastritis is a significant risk factor for developing gastric cancer. The incorporation of machine learning algorithms can efficiently elevate the possibility of accurately detecting atrophic gastritis. Nevertheless, when the trained model is applied in real-life circumstances, its output is often not consistently reliable. In this paper, we propose Adaptify, an adaptation scheme in which the model assimilates knowledge from its own classification decisions. Our proposed approach includes keeping the primary model constant, while simultaneously running and updating the auxiliary model. By integrating the knowledge gleaned by the auxiliary model into the primary model and merging their outputs, we have observed a notable improvement in output stability and consistency compared to relying solely on either the main model or the auxiliary model.