Asad Safi

Computer---Aided diagnosis of pigmented skin dermoscopic images

Diagnosis of benign and malign skin lesions is currently mostly relying on visual assessment and frequent biopsies performed by dermatologists. As the timely and correct diagnosis of these skin lesions is one of the most important factors in the therapeutic outcome, leveraging new technologies to assist the dermatologist seems natural. In this paper we propose a machine learning approach to classify melanocytic lesions into malignant and benign from dermoscopic images. The dermoscopic image database is composed of 4240 benign lesions and 232 malignant melanoma. For segmentation we are using multiphase soft segmentation with total variation and H1 regularization. Then, each lesion is characterized by a feature vector that contains shape, color and texture information, as well as local and global parameters that try to reflect structures used in medical diagnosis. The learning and classification stage is performed using SVM with polynomial kernels. The classification delivered accuracy of 98.57% with a true positive rate of 0.991% and a false positive rate of 0.019%.

Detection and Counting of On-Tree Citrus Fruit for Crop Yield Estimation

In this paper, we present a technique to estimate citrus fruit yield from the tree images. Manually counting the fruit for yield estimation for marketing and other managerial tasks is time consuming and requires human resources, which do not always come cheap. Different approaches have been used for the said purpose, yet separation of fruit from its background poses challenges, and renders the exercise inaccurate. In this paper, we use k-means segmentation for recognition of fruit, which segments the image accurately thus enabling more accurate yield estimation. We created a dataset containing 83 tree images with 4001 citrus fruits from three different fields. We are able to detect the on-tree fruits with an accuracy of 91.3%. In addition, we find a strong correlation between the manual and the automated fruit count by getting coefficients of determination R2 up to 0.99.

Pedestrian detection using HOG, LUV and optical flow as features with AdaBoost as classifier

Pedestrian detection has been used in applications such as car safety, video surveillance, and intelligent vehicles. In this paper, we present a pedestrian detection scheme using HOG, LUV and optical flow features with AdaBoost Decision Stump classifier. Our experiments on Caltech-USA pedestrian dataset show that the proposed scheme achieves promising results of about 16.7% log-average miss rate.

Manifold learning for dimensionality reduction and clustering of skin spectroscopy data

Diagnosis of benign and malign skin lesions is currently done mostly relying on visual assessment and frequent biopsies performed by dermatologists. As the timely and correct diagnosis of these skin lesions is one of the most important factors in the therapeutic outcome, leveraging new technologies to assist the dermatologist seems natural. Optical spectroscopy is a technology that is being established to aid skin lesion diagnosis, as the multi-spectral nature of this imaging method allows to detect multiple physiological changes like those associated with increased vasculature, cellular structure, oxygen consumption or edema in tumors. However, spectroscopy data is typically very high dimensional (on the order of thousands), which causes difficulties in visualization and classification. In this work we apply different manifold learning techniques to reduce the dimensions of the input data and get clustering results. Spectroscopic data of 48 patients with suspicious and actually malignant lesions was analyzed using ISOMAP, Laplacian Eigenmaps and Diffusion Maps with varying parameters and compared to results using PCA. Using optimal parameters, both ISOMAP and Laplacian Eigenmaps could cluster the data into suspicious and malignant with 96% accuracy, compared to the diagnosis of the treating physicians.

Skin lesions classification with optical spectroscopy

Diagnosis of benign and malign skin lesions is currently mostly relying on visual assessment and frequent biopsies performed by dermatologists. As the timely and correct diagnosis of these skin lesions is one of the most important factors in the therapeutic outcome, leveraging new technologies to assist the dermatologist seems natural. Complicating matters is a blood cancer called Cutaneous T-Cell Lymphoma, which also exhibits symptoms as skin lesions. We propose a framework using optical spectroscopy and a multi-spectral classification scheme using support vector machines to assist dermatologists in diagnosis of normal, benign and malign skin lesions. As a first step we show successful classification (94.9%) of skin lesions from regular skin in 48 patients based on 436 measurements. This forms the basis for future automated classification of different skin lesions in diseased patients.

Vehicle Remote Health Monitoring and Prognostic Maintenance System

In many industries inclusive of automotive vehicle industry, predictive maintenance has become more important. It is hard to diagnose failure in advance in the vehicle industry because of the limited availability of sensors and some of the designing exertions. However with the great development in automotive industry, it looks feasible today to analyze sensor’s data along with machine learning techniques for failure prediction. In this article, an approach is presented for fault prediction of four main subsystems of vehicle, fuel system, ignition system, exhaust system, and cooling system. Sensor is collected when vehicle is on the move, both in faulty condition (when any failure in specific system has occurred) and in normal condition. The data is transmitted to the server which analyzes the data. Interesting patterns are learned using four classifiers, Decision Tree, Support Vector Machine, Nearest Neighbor, and Random Forest. These patterns are later used to detect future failures in other vehicles which show the similar behavior. The approach is produced with the end goal of expanding vehicle up-time and was demonstrated on 70 vehicles of Toyota Corolla type. Accuracy comparison of all classifiers is performed on the basis of Receiver Operating Characteristics (ROC) curves.

Outdoor Scene Labeling Using ALE and LSC Superpixels

Scene labeling has been an important and popular area of computer vision and image processing for the past few years. It is the process of assigning pixels to specific predefined categories in an image. A number of techniques have been proposed for scene labeling but all have some limitations regarding accuracy and computational time. Some methods only incorporate the local context of images and ignore the global information of objects in an image. Therefore, accuracy of scene labeling is low for these methods. There is a need to address these issues of scene labeling to improve labeling accuracy. In this paper, we perform outdoor scene labeling using Automatic labeling Environment (ALE). We enhance this framework by incorporating bilateral filter based preprocessing, LSC superpixels and large co-occurrence weight. Experiments on a publicly available MSRC v1 dataset showed promising results with 89.44% pixel-wise accuracy and 78.02% class-wise accuracy.

Online Breast Cancer Diagnosis System

Breast cancer is the most common cancer among women, and is the second leading cause of death after lung cancer. The ability to accurately identify the malignancy in early stage is the key for better prognosis and preparation of effective treatment. In the developing world, even though at times imaging machines are available in the rural areas but due to the absence of the relevant medical expertise early detection of cancer remains only a pipe dream. Along with the imaging machines, internet has made inroads into rural surroundings. That makes the availability of online automatic systems that can identify the presence or absence of malignancy, without human involvement, an important aspect of healthcare systems in the underdeveloped rural surroundings. This paper presents an online tumor detection application, that uses mammogram images. The mammogram images taken at a local facility are transferred over the internet to a remote server that hosts the application that can classify tumour. It was trained on 322 mammographic images, from the mini-MIAS database. We have achieved sensitivity of 90.15 %.

Feasibility Study of Optical Spectroscopy as a Medical Tool for Diagnosis of Skin Lesions

Skin cancer is one of the most frequently en-countered types of cancer in the Western world. According to the Skin Cancer Foundation Statistics, one in every five Americans develops skin cancer during his/her lifetime. Today, the incurability of advanced cutaneous melanoma raises the importance of its early detection. Since the differentiation of early melanoma from other pigmented skin lesions is not a trivial task, even for experienced dermatologists, computer aided diagnosis could become an important tool for reducing the mortality rate of this highly malignant cancer type. In this paper, a computer aided diagnosis system based on machine learning is proposed in order to support the clinical use of optical spectroscopy for skin lesions quantification and classification. The focuses is on a feasibility study of optical spectroscopy as a medical tool for diagnosis. To this end, data acquisition protocols for optical spectroscopy are defined and detailed analysis of feature vectors is performed. Different tech-niques for supervised and unsupervised learning are explored on clinical data, collected from patients with malignant and benign skin lesions.

Intelligent Pedestrian Detection using Optical Flow and HOG

Pedestrian detection is an important aspect of autonomous vehicle driving as recognizing pedestrians helps in reducing accidents between the vehicles and the pedestrians. In literature, feature based approaches have been mostly used for pedestrian detection. Features from different body portions are extracted and analyzed for interpreting the presence or absence of a person in a particular region in front of car. But these approaches alone are not enough to differentiate humans from non-humans in dynamic environments, where background is continuously changing. We present an automated pedestrian detection system by finding pedestrians’ motion patterns and combing them with HOG features. The proposed scheme achieved 17.7% and 14.22% average miss rate on ETHZ and Caltech datasets, respectively.