Automated White Blood Cell Disease Recognition Using Lightweight Deep Learning

White blood cells (WBC) are immune system cells, which is why they are also known as immune cells. They protect the human body from a variety of dangerous diseases and outside invaders. The majority of WBCs come from red bone marrow, although some come from other important organs in the body. Because manual diagnosis of blood disorders is difficult, it is necessary to design a computerized technique. Researchers have introduced various automated strategies in recent years, but they still face several obstacles, such as imbalanced datasets, incorrect feature selection, and incorrect deep model selection. We proposed an automated deep learning approach for classifying white blood disorders in this paper. The data augmentation approach is initially used to increase the size of a dataset. Then, a Darknet-53 pre-trained deep learning model is used and fine-tuned according to the nature of the chosen dataset. On the fine-tuned model, transfer learning is used, and features engineering is done on the global average pooling layer. The retrieved characteristics are subsequently improved with a specified number of iterations using a hybrid reformed binary grey wolf optimization technique. Following that, machine learning classifiers are used to classify the selected best features for final classification. The experiment was carried out using a dataset of increased blood diseases imaging and resulted in an improved accuracy of over 99%.     

A Particle Swarm Optimization Based Deep Learning Model for Vehicle Classification

Image classification is a core field in the research area of image processing and computer vision in which vehicle classification is a critical domain. The purpose of vehicle categorization is to formulate a compact system to assist in real-world problems and applications such as security, traffic analysis, and self-driving and autonomous vehicles. The recent revolution in the field of machine learning and artificial intelligence has provided an immense amount of support for image processing related problems and has overtaken the conventional, and handcrafted means of solving image analysis problems. In this paper, a combination of pre-trained CNN GoogleNet and a nature-inspired problem optimization scheme, particle swarm optimization (PSO), was employed for autonomous vehicle classification. The model was trained on a vehicle image dataset obtained from Kaggle that has been suitably augmented. The trained model was classified using several classifiers; however, the Cubic SVM (CSVM) classifier was found to outperform the others in both time consumption and accuracy (94.8%). The results obtained from empirical evaluations and statistical tests reveal that the model itself has shown to outperform the other related models not only in terms of accuracy (94.8%) but also in terms of training time (82.7 s) and speed prediction (380 obs/sec).     

Multimodal brain tumor detection and classification using deep saliency map and improved dragonfly optimization algorithm

In the last decade, there has been a significant increase in medical cases involving brain tumors. Brain tumor is the tenth most common type of tumor, affecting millions of people. However, if it is detected early, the cure rate can increase. Computer vision researchers are working to develop sophisticated techniques for detecting and classifying brain tumors. MRI scans are primarily used for tumor analysis. We proposed an automated system for brain tumor detection and classification using a saliency map and deep learning feature optimization in this paper. The proposed framework was implemented in stages. In the initial phase of the proposed framework, a fusion-based contrast enhancement technique is proposed. In the following phase, a tumor segmentation technique based on saliency maps is proposed, which is then mapped on original images based on active contour. Following that, a pre-trained CNN model named EfficientNetB0 is fine-tuned and trained in two ways: on enhanced images and on tumor localization images. Deep transfer learning is used to train both models, and features are extracted from the average pooling layer. The deep learning features are then fused using an improved fusion approach known as Entropy Serial Fusion. The best features are chosen in the final step using an improved dragonfly optimization algorithm. Finally, the best features are classified using an extreme learning machine (ELM). The experimental process is conducted on three publically available datasets and achieved an improved accuracy of 95.14, 94.89, and 95.94%, respectively. The comparison with several neural nets shows the improvement of proposed framework.     

Reinforcement Learning with an Ensemble of Binary Action Deep Q-Networks

With the advent of Reinforcement Learning (RL) and its continuous progress, state-of-the-art RL systems have come up for many challenging and real-world tasks. Given the scope of this area, various techniques are found in the literature. One such notable technique, Multiple Deep Q-Network (DQN) based RL systems use multiple DQN-based-entities, which learn together and communicate with each other. The learning has to be distributed wisely among all entities in such a scheme and the inter-entity communication protocol has to be carefully designed. As more complex DQNs come to the fore, the overall complexity of these multi-entity systems has increased many folds leading to issues like difficulty in training, need for high resources, more training time, and difficulty in fine-tuning leading to performance issues. Taking a cue from the parallel processing found in the nature and its efficacy, we propose a lightweight ensemble based approach for solving the core RL tasks. It uses multiple binary action DQNs having shared state and reward. The benefits of the proposed approach are overall simplicity, faster convergence and better performance compared to conventional DQN based approaches. The approach can potentially be extended to any type of DQN by forming its ensemble. Conducting extensive experimentation, promising results are obtained using the proposed ensemble approach on OpenAI Gym tasks, and Atari 2600 games as compared to recent techniques. The proposed approach gives a stateof-the-art score of 500 on the Cartpole-v1 task, 259.2 on the LunarLander-v2 task, and state-of-the-art results on four out of five Atari 2600 games.