Epilepsy is a condition that disrupts normal brain function and sometimes leads to seizures, unusual sensations, and temporary loss of awareness. Electroencephalograph (EEG) records are commonly used for diagnosing epilepsy, but traditional analysis is subjective and prone to misclassification. Previous studies applied Deep Learning (DL) techniques to improve EEG classification, but their performance has been limited due to dynamic and non-stationary nature of EEG structure. In this paper, we propose a multi-channel EEG classification model called LConvNet, which combines Convolutional Neural Networks (CNN) for spatial feature extraction and Long Short-Term Memory (LSTM) for capturing temporal dependencies. The model is trained using open source secondary EEG data from Temple University Hospital (TUH) to distinguish between epileptic and healthy EEG signals. Our model achieved an impressive accuracy of 97%, surpassing existing EEG classification models used in similar tasks such as EEGNet, DeepConvNet and ShallowConvNet that had 86%, 96% and 78% respectively. Furthermore, our model demonstrated impressive performance in terms of trainability, scalability and parameter efficiency during additional evaluations.

Abstract A Fuzzy logic based mean filter (FLBMF) is presented for impulse noise reduction of mammogram images degraded with additive impulse noise. FLBMF removes both low and high density impulsive noise from mammogram images. FLBMF performs this in three major phases. In phase one, the detection of noisy pixels is performed and determined. In phase two, an adaptive threshold is determined by examining the neighboring pixels. In phase three, fuzzy membership functions and fuzzy rules are used to decide whether the current pixel is noise-free, or the noise pixel is in a smooth or detailed region. All these phases are based on fuzzy rules making use of membership functions. FLBMF can be applied iteratively to effectively reduce impulsive noise. In particular, the membership function’s shape is adapted according to the remaining noise level after each iteration, making use of the distribution of the homogeneity in the image. In this approach, the mammogram images are selected from mini-MIAS database and renamed as MammoB1, MammoB2, MammoB3 and MammoB4, are then deformed by varying intensities of impulse noise. The performance evaluation of various filters including FLBMF tested at low, medium and high noise densities on different standard grey scale mammogram images is then carried out. Mathematical performance parameters including Mean Square Error (MSE), Peak-signal-to-noise-ratio (PSNR), and Structural Similarity Index Measure (SSIM) are finally applied to measure the accuracy and performance of this approach. The image modalities implementation and analysis of our approach is carried out in MATLAB functions. Keywords: Impulsive Noise; FLBMF, Fuzzy membership function, Fuzzy rules, Edge preserving filtering, Fuzzy image filtering, Noise reduction

Adapting the target dataset for a pre-trained model is still challenging. These adaptation problems result from a lack of adequate transfer of traits from the source dataset; this often leads to poor model performance resulting in trial and error in selecting the best performing pre-trained model. This paper introduces the conflation of source domain low-level textural features extracted using the first layer of the pretrained model. The extracted features are compared to the conflated low-level features of the target dataset to select a higher quality target dataset for improved pre-trained model performance and adaptation. From comparing the various probability distance metrics, Kullback-Leibler is adopted to compare the samples from both domains. We experiment on three publicly available datasets and two ImageNet pre-trained models used in past studies for results comparisons. This proposed approach method yields two categories of the target samples with those with lower Kullback-Leibler values giving better accuracy, precision and recall. The samples with the lower Kullback-Leibler values give a higher margin accuracy rate of 6.21% to 7.27%, thereby leading to better model adaptation for target transfer learning datasets and tasks

The selection of layers in the transfer learning fine-tuning process ensures a pre-trained model’s accuracy and adaptation in a new target domain. However, the selection process is still manual and without clearly defined criteria. If the wrong layers in a neural network are selected and used, it could lead to poor accuracy and model generalisation in the target domain. This paper introduces the use of Kullback-Leibler divergence on the weight correlations of the model’s convolutional neural network layers. The approach identifies the positive and negative weights in the ImageNet initial weights selecting the best-suited layers of the network depending on the correlation divergence. We experiment on four publicly available datasets and four ImageNet pre-trained models that have been used in past studies for results comparisons. This proposed approach method yields better accuracies than the standard fine-tuning baselines with a margin accuracy rate of 10.8% to 24%, thereby leading to better model adaptation for target transfer learning tasks.

Social media has been embraced by different people as a convenient and official medium of communication. People write messages and attach images and videos on Twitter, Facebook and other social media which they share. Social media therefore generates a lot of data that is rich in sentiments from these updates. Sentiment analysis has been used to determine opinions of clients, for instance, relating to a particular product or company. Knowledge based approach and Machine learning approach are among the strategies that have been used to analyze these sentiments. The performance of sentiment analysis is however distorted by noise, the curse of dimensionality, the data domains and size of data used for training and testing. This research aims at developing a model for sentiment analysis in which dimensionality reduction and the use of different parts of speech improves sentiment analysis performance. It uses natural language processing for filtering, storing and performing sentiment analysis on the data from social media. The model is tested using Naïve Bayes, Support Vector Machines and K-Nearest neighbor machine learning algorithms and its performance compared with that of two other Sentiment Analysis models. Experimental results show that the model improves sentiment analysis performance using machine learning techniques.