Leukocytes identification using augmentation and transfer learning based convolution neural network

Abstract

Most haematological diseases can be diagnosed using the morphological analysis of the microscopic blood image. The basic routine of the morphological analysis can be performed using the microscopic device which requires the skills and experiences of the haematologists. An inexperienced haematologist can lead to critical human errors. Therefore, this paper aims to propose an automated classification system used to classify different types of leukocytes based on the convolution neural network (CNN) algorithm. CNN has achieved robust performance in various fields especially in medical applications. A dataset of microscopic blood cells images of the conforming tags (basophil, eosinophil, erythroblast, lymphocyte, monocyte, neutrophil, and platelet) was used to train and test the proposed algorithm. The augmentation and deep transfer approaches were used to improve and enhance the performance of the CNN algorithm. The overall accuracy of the proposed classifier was 98% with Visual Geometry Group-19 (VGG-19). The obtained accuracy was higher than the state-of-art algorithms. To conclude that using the augmentation and deep transfer approaches with VGG-19 can obtain better classification results.

Handwritten Javanese script recognition method based 12-layers deep convolutional neural network and data augmentation

“…Data augmentation is carried out to ensure the datasets vary. Besides, it is conducted to improve the classification accuracy performance [17], [25]- [29]. Figure 2 shows the method proposed in this research, further described in detail in subsections 3.1 to 3.…”

Image classification-based transfer learning framework for image detection of IoT devices

“…In [15]- [18], CNN-based TL methods are implemented for COVID-19 detection using chest X-ray images. The studies in [19]- [22] also apply CNN-based TL techniques to identify leukocytes as well as red blood cells for blood-related diseases and classify fundus for general retinal diseases diagnosis, respectively. The TL with pre-trained CNN models are also used for non-medical applications including Thai culture and Pitha traditional food images classification [23], [24], and other applications including land cover, fabric defect, birds' species, distracted driver classifications as well as age-invariant face recognition [25]- [29].…”