Search

Background:We describe the development of a new mobile app called “FootSnap,” to standardize photographs of diabetic feet and test its reliability on different occasions and between different opera...

This paper presents an automated facial wrinkles annotator for coarse wrinkles, fine wrinkles and wrinkle depth map extraction. First we extended Hybrid Hessian Filter by introducing a multi-scale filter to isolate the coarse wrinkles from fine wrinkles. Then we generate a wrinkle probabilistic map. When evaluated on 20 high resolution full face images (10 from our in-house dataset and 10 from FERET dataset), we achieved good accuracy when the result of coarse wrinkles was validated with manual annotation. Furthermore, we visually illustrate the ability of the annotator in detecting fine wrinkles. This paper advances the field by automate the localisation of the fine wrinkles, which might not be possible to annotate manually. Our automated facial wrinkles annotator will be beneficial to large-scale data annotation and cosmetic applications.

Anomaly detection on road traffic is an important task due to its great potential in urban traffic management and road safety. It is also a very challenging task since the abnormal event happens very rarely and exhibits different behaviors. In this work, we present a model to detect anomaly in road traffic by learning from the vehicle motion patterns in two distinctive yet correlated modes, i.e., the static mode and the dynamic mode, of the vehicles. The static mode analysis of the vehicles is learned from the background modeling followed by vehicle detection procedure to find the abnormal vehicles that keep still on the road. The dynamic mode analysis of the vehicles is learned from detected and tracked vehicle trajectories to find the abnormal trajectory which is aberrant from the dominant motion patterns. The results from the dual-mode analyses are finally fused together by driven a re-identification model to obtain the final anomaly. Experimental results on the Track 2 testing set of NVIDIA AI CITY CHALLENGE show the effectiveness of the proposed dual-mode learning model and its robustness in different real scenes. Our result ranks the first place on the final Leaderboard of the Track 2.

Wrinkles can be embedded in several image-based applications as a descriptor for human skin. However, wrinkle-based age estimation research has not been widely addressed. In this paper, we introduce a Multi-scale Wrinkle Patterns (MWP) representation, investigate the effect of wrinkles on face age estimation and propose Hybrid Ageing Patterns (HAP) for face age estimation. To define the wrinkle regions more precisely, a template consisting of 10 regions constructed relatively to a set of automatically located facial landmarks is used. We extract the multi-scale wrinkles in each region and encode them into MWP. We use Support Vector Regression to estimate age from the combination of such patterns. The performance of the algorithms is assessed by using Mean Absolute Error (MAE) on three state-of-the-art datasets - FG-NET, FERET and MORPH. We observe that MWP produces a comparable MAE of 4.16 on FERET to the state of the art. Finally we propose HAP, which combines the features from MWP and the facial appearance model (FAM), and demonstrate improved performance on FERET and MORPH with MAE of 3.02 (±2.92) and 3.68 (±2.98), respectively. Therefore, we conclude that MWP is an important complementary feature for face age estimation.

We propose the Component Bio-Inspired Feature (CBIF) with a moving segmentation scheme for age estimation. The CBIF defines a superset for the commonly used Bio-Inspired Feature (BIF) with more parameters and flexibility in settings, resulting in features with abundant characteristics. An in-depth study is performed for the determination of the parameters good for capturing age-related traits. The moving segmentation is proposed to better determine the age boundaries good for age grouping, and improve the overall performance. The proposed approach is evaluated on two common benchmarks, FG-NET and MORPH databases, and compared with contemporary approaches to demonstrate its efficacy.

Face features convey many personal information that promote and regulate our social linkages. Age prediction using single layer estimation such as aging subspace or a hybrid pattern is limited due to the complexity of human faces. In this work, we propose Multilayer Age Regression (MAR) where the face age is predicted based on a coarse-to-fine estimation using global and local features. In the first layer, Support Vector Regression (SVR) performs a between group prediction by the parameters of Facial Appearance Model (FAM). In the second layer, a within group estimation is performed using FAM, Bio-Inspired Features (BIF), Kernel-based Local Binary Patterns (KLBP) and Multi-scale Wrinkle Patterns (MWP). The performance of MAR is assessed on four benchmark datasets: FGNET, MORPH, FERET and PAL. Results showed that MAR outperforms the state of the art on FERET with a Mean Absolute Error (MAE) of 3.00 (±4.14).

Micro-facial expressions are regarded as an important human behavioural event that can highlight emotional deception. Spotting these movements is difficult for humans and machines, however research into using computer vision to detect subtle facial expressions is growing in popularity. This paper proposes an individualised baseline micro-movement detection method using 3D Histogram of Oriented Gradients (3D HOG) temporal difference method. We define a face template consisting of 26 regions based on the Facial Action Coding System (FACS). We extract the temporal features of each region using 3D HOG. Then, we use Chi-square distance to find subtle facial motion in the local regions. Finally, an automatic peak detector is used to detect micro-movements above the newly proposed adaptive baseline threshold. The performance is validated on two FACS coded datasets: SAMM and CASME II. This objective method focuses on the movement of the 26 face regions. When comparing with the ground truth, the best result was an AUC of 0.7512 and 0.7261 on SAMM and CASME II, respectively. The results show that 3D HOG outperformed for micro-movement detection, compared to state-of-the-art feature representations: Local Binary Patterns in Three Orthogonal Planes and Histograms of Oriented Optical Flow.

The majority of current facial age estimation methods are based on appearance based features. However, wrinkle based research has not been widely addressed. In this paper, we propose a novel method based on multi-scale aging patterns (MAP). These directly extract the features from local patches without extensive geometric modelling. First, we locate facial landmarks by using the Face++ detector and then normalize the face by using a linear transformation. We define a face template which consists of ten predefined wrinkle regions. Then, for each region, we detect wrinkles and construct aging patterns by using the MAP. Finally, the age is estimated by implementing the sequential minimal optimization (SMO). The performance of the algorithms is assessed by using mean absolute error (MAE) on the benchmark database - FERET. We observe that MAP produces a lower MAE of 4.87 on FERET compared to the benchmark algorithms. Therefore, we conclude that wrinkle could be used as a feature on face age estimation. Future work would involve improvements of the algorithm by combining other descriptors such as non-wrinkle descriptor and appearance parameters.

Aging as a natural phenomenon affects different parts of the human body under the influence of various biological and environmental factors. The most pronounced changes that occur on the face is the appearance of wrinkles, which are the focus of this research. Accurate wrinkle detection is an important task in face analysis. Some have been proposed in the literature, but the poor localization limits the performance of wrinkle detection. It will lead to false wrinkle detection and consequently affect the processes such as age estimation and clinician score assessment. Therefore, we propose a hybrid Hessian filter (HHF) to cope with the identified problem. HHF is composed of the directional gradient and Hessian matrix. The proposed filter is conceptually simple, however, it significantly increases the true wrinkle localization when compared with the conventional methods. In the experimental setup, three coders have been instructed to annotate the wrinkle of 2D forehead image manually. The inter-reliability among three coders is 93 % of Jaccard similarity index (JSI). In comparison to the state-of-the-art Cula method (CLM) and Frangi filter, HHF yielded the best result with a mean JSI of 75.67 %. We noticed that the proposed method is capable of detecting the medium to coarse wrinkle but not the fine wrinkle. Although there is a gap between human annotation and automated detection, this work demonstrates that HHF is a remarkably strong filter for wrinkle detection. From the experimental results, we believe that our findings are notable in terms of the JSI.

Diabetic foot ulcers (DFUs) and other foot pathologies are highly prevalent in people with diabetes, especially in patients with the secondary complications of peripheral neuropathy and/or peripheral vascular disease.1 The development of DFUs involves changes in the appearance of the foot over time, however, many subtle changes in foot shape, texture, and color will likely be undetectable by human vision until a DFU is established. Advanced computer vision algorithms have been applied in the medical field to detect subtle signs to differentiate the type of tumours2 subtle changes in facial features—such as fine wrinkles on the face3 and micro facial movements4—where some of these features are not detectable by human vision. We see the opportunity to apply computer vision algorithms to help in the early detection of key pathological changes in the diabetic foot leading to the development of a DFU.

Natural language processing is an emerging technology in linguistic industry and an aid to human-computer interaction in computer science. Language identification, on the other hand, is a form of pattern recognition that helps to identify predefined language of a web page and to predict the unknown language of one particular text. Written texts are constructed by common features such as character, word and n-gram and these characteristics are unique among languages. From the experiment result, the performance of the supervised n-gram produces an accurate identification value and outperforms the distance measurement on Arabic script web pages.

Language is an indispensable tool for human communication, and presently, the language that dominates the Internet is English. Language identification is the process of determining a predetermined language automatically from a given content (e.g., English, Malay, Danish, Estonian, Czech, Slovak, etc.). The ability to identify other languages in relation to English is highly desirable. It is the goal of this research to improve the method used to achieve this end. Three methods have been studied in this research are distance measurement, Boolean method, and the proposed method, namely, optimum profile. From the initial experiments, we have found that, distance measurement and Boolean method is not reliable in the European web page identification. Therefore, we propose optimum profile which is using N-grams frequency and N-grams position to do web page language identification. The result show that the proposed method gives the highest performance with accuracy 91.52%.

Language identification is the process identifying predefined language in a document automatically; we focused on the web documents in this paper. Initially, we have applied the letter frequency as features combine with neural networks in Arabic script language identification. However, reliability of selected letters of the features is a major issue to be overcome. Therefore, we propose an improved letter weighting feature selection in order to enhance the effectiveness of language identification. It is based on the concept letter frequency document frequency. From the experiments, we have found that the improved letter weighting feature selection achieve the highest accuracy 99.75% on Arabic script language identification.

Current techniques of language identification are based on a method that assigns one or more textual documents into a set of predefined languages that are relevant to page contents. In this paper, we proposed an agent architecture that is used in criminal mobile devices identification systems. It is based on the usage of a software agent to process at least one document by using a dictionary that belong to a set of languages in order to determine the type of language features to be used in the text preprocessing module. Then, the agent will map at least one of the documents with the content of the dictionary in order to identify the languages used in the text by the language identification agent. Finally, the digital forensic agent will check the potential criminal short messages through the predefined keyword repository of corresponding language. Form our experiments, the agent architecture has been able to identify correctly the types of languages written in the short text messaging (SMS) system.

In general, a web document page may contain several script forms. Each script can be used for constructing different languages. Determining the languages of the document is the required to effectively be able to apply many search and information retrieval techniques. In this work, we propose hybrid-grams feature selection methods by integrating unigram and bigrams. The method makes use of local statistical information or data within a document to determine the language. From the experiments, we have noticed that hybrid-grams are outperformed than unigram and bigrams in Cyrillic and Indic script language identifications.

The information available in languages other than English on the World Wide Web and global information systems is increasing significantly. Different languages can be produced by using one particular script such as Arabic, Persian, Urdu and Pashto that use Arabic script letters. The issue is how to produce reliable features of a web page that is to undergo language identification. Incorrectly identifying the language results in garbled translations as well as faulty and incomplete analyses. The aim of this study is to enhance the effectiveness of feature selection method of web page language identification. We have investigated total N-grams, N-grams frequency, N-grams frequency document frequency, and N-grams frequency inverse document frequency of web page language identification. From the experimental results, it is proven that N-grams frequency gives the most promising result compared to other feature selection methods.