Ultrafast-and-Ultralight ConvNet-Based Intelligent Monitoring System for Diagnosing Early-Stage Mpox Anytime and Anywhere

TL;DR

An ultrafast and ultralight network named Fast‐MpoxNet, utilizing transfer learning and data augmentation, which has the potential to mitigate future mpox outbreaks and pave the way for developing real‐time diagnostic tools in the healthcare field.

Abstract

Due to the absence of more efficient diagnostic tools, the spread of mpox continues to be unchecked. Although related studies have demonstrated the high efficiency of deep learning models in diagnosing mpox, key aspects such as model inference speed and parameter size have always been overlooked. Herein, an ultrafast and ultralight network named Fast-MpoxNet is proposed. Fast-MpoxNet, with only 0.27M parameters, can process input images at 68 frames per second (FPS) on the CPU. To detect subtle image differences and optimize model parameters better, Fast-MpoxNet incorporates an attention-based feature fusion module and a multiple auxiliary losses enhancement strategy. Experimental results indicate that Fast-MpoxNet, utilizing transfer learning and data augmentation, produces 98.40% classification accuracy for four classes on the mpox dataset. Furthermore, its Recall for early-stage mpox is 93.65%. Most importantly, an application system named Mpox-AISM V2 is developed, suitable for both personal computers and smartphones. Mpox-AISM V2 can rapidly and accurately diagnose mpox and can be easily deployed in various scenarios to offer the public real-time mpox diagnosis services. This work has the potential to mitigate future mpox outbreaks and pave the way for developing real-time diagnostic tools in the healthcare field.

Figures (11)

• Figure 1: a: Monkeypox samples, b: Chickenpox samples, c: Measles samples, d: Normal samples.
• Figure 2: Some samples of augmented dataset and comparison between original images and augmented images. a: Monkeypox samples, b: Chickenpox samples, c: Measles samples.
• Figure 3: a: The overall architecture of Fast-MpoxNet. Here, PWConv is Point-wise Convolution which can change the number of channels of the feature map and has the characteristics of low computational power consumption and low memory access frequency. BN represents Batch Normalization. Besides, C×H×W represents the number of channels, image height, and image width. b: Internal structure of Cls Head. c: Internal structure of Attention-Based Local and Global Fusion Module (ABLGFM).
• Figure 4: The workflow of our study.
• Figure 5: a: The Accuracy between four models on each fold. b: The average Accuracy of different models in the ablation experiment. c: The Precision between four models for various diseases. d: The Recall between four models for various diseases. e: The Specificity between four models for various diseases. f: The F1-score between four models for various diseases.