Skin lesion segmentation and classification: A unified framework of deep neural network features fusion and selection

Abstract

Automated skin lesion diagnosis from dermoscopic images is a difficult process due to several notable problems such as artefacts (hairs), irregularity, lesion shape, and irrelevant features extraction. These problems make the segmentation and classification process difficult. In this research, we proposed an optimized colour feature (OCF) of lesion segmentation and deep convolutional neural network (DCNN)‐based skin lesion classification. A hybrid technique is proposed to remove the artefacts and improve the lesion contrast. Then, colour segmentation technique is presented known as OCFs. The OCF approach is further improved by an existing saliency approach, which is fused by a novel pixel‐based method. A DCNN‐9 model is implemented to extract deep features and fused with OCFs by a novel parallel fusion approach. After this, a normal distribution‐based high‐ranking feature selection technique is utilized to select the most robust features for classification. The suggested method is evaluated on ISBI series (2016, 2017, and 2018) datasets. The experiments are performed in two steps and achieved average segmentation accuracy of more than 90% on selected datasets. Moreover, the achieve classification accuracy of 92.1%, 96.5%, and 85.1%, respectively, on all three datasets shows that the presented method has remarkable performance.

Classification of stomach infections: A paradigm of convolutional neural network along with classical features fusion and selection

“…Features fusion is an important step of PR field in which different features vectors are concatenated to obtain final feature vector for classification (Khan, Akram, et al, ; Khan, Lali, et al, ; Khan, Javed, Sharif, Saba, & Rehman, ; Khan, Rashid, Sharif, Javed, & Akram, ; Khan, Sharif, Akram, Bukhari, & Nayak, ; Khan, Sharif, Raza, Anjum, et al, ). Main motivation of this step is to put all the information of different descriptors in a single feature vector which may be helpful for minimum error rate.…”

Malignant Melanoma Classification Using Deep Learning: Datasets, Performance Measurements, Challenges and Opportunities

“…Performance can further be enhanced if the proposed method uses autoencoder for feature selection and classification. Attique et al [59] incorporated optimized color feature segmentation with CNN, and it uses ISBI 2016, ISBI 2017 and ISBI 2018 datasets to check the effectiveness of its proposed method. El-Khatiab et al [81] use transfer learning which was based on Google Net, ResNet and NasNet.…”

“…The research articles [47], [48], [54], [55], [59], [61], [64], [68], [73], [75] and [78] In this review [67], [70], [72], [76] and [94] use this dataset to check the accuracy of their methods e) ISBI 2017 Dataset Challenge [37]: This dataset was used by [47], [50], [58], [59], [63], [64], [71] and [77] for testing their proposed methods. It contained 2,000 dermoscopic images in which 374 were melanomas, 254 were seborrheic keratoses, and 1,372 benign nevi images.…”

“…g) 2018 ISIC challenge [39]: This dataset contained more than 10,000 dermoscopic images of 7 types of diseases (melanoma, nevi, seborrheic keratosis, BCC, Bowen's disease and actinic keratosis, vascular lesions, and dermatofibromes). The studies include [48], [59], [60] and [69]…”

Deep Learning based Semantic Similarity Detection using Text Data

“…Due to the challenges and need, text classification has high demand. Recently, deep learning shows a huge advancement in many machine learning research areas such as text/signature [6], object classification [21,47], visual surveillance [26,51], biometrics [3], medical imaging [27,29,42,48,52], and many more [4,25,28,[30][31][32]41]. In this research, a deep learning model is proposed for detecting similarity between short sentences such as question pairs.…”