Wooi-Haw Tan

Impulse Detection Adaptive Fuzzy (IDAF) Filter

An Impulse Detection Adaptive Fuzzy (IDAF) filter is proposed in this paper, to achieve improved filtering of impulsive noise while preserving image details. It is a spatial filter which operates in 2-stage 3x3 windows where the update value of the central pixel is a function of the median value of the pixels in the window. The proposed IDAF filter operates as follows. First, an impulse detection method classifies each pixel to be noisy pixel or image pixel. Then, if a pixel is assumed to be noisy, it is not used for deciding the new value of other pixels. A scheme is introduced to obtain a good replacement pixel value, which is then stored. The median of the stored values is a considered a good estimate of the image pixel. Finally, the adaptive fuzzy filter will then assign weights to the stored pixel values to produce the central pixel’s new value. Weights are assigned to remove impulse noise or preserve the image details, depending on the pixel’s characteristics. These characteristics are identified from the noisy image beforehand using a compatibility measure. The process detailed above enables the IDAF filter to remove impulse noise while preserving image details. Through testing, our experimental results show that the IDAF filter performs better than other standard median based methods.

Extraction of human gait features from enhanced human silhouette images

In this paper, a new approach is proposed for extracting human gait features from a walking human based on the silhouette image. The approach consists of five stages: clearing the background noise of image by morphological opening; measuring the width and height of the human silhouette; dividing the enhanced human silhouette into six body segments based on anatomical knowledge; applying morphological skeleton to obtain the body skeleton; and applying Hough transform to obtain the joint angles from the body segment skeletons. The joint angles together with the height and width of the human silhouette are collected and used for gait analysis. From the experiment conducted, it can be observed that the proposed system is feasible as satisfactory results have been achieved.

Uncommitted Morphological Merging of Watershed Segments

Over the past decades, watershed segmentation has gained much popularity due to its attractive properties: closer to what a human observer would decide, efficient implementation and elegant combination of edge- and region-based segmentation. Nevertheless, it is often associated with the problem of over-segmentation owing to its high sensitivity to variation in intensity. Various merging algorithms have been proposed to tackle this problem, but most of them are either user-dependent, knowledge-based, or computationally expensive. This paper presents a novel merging algorithm, which is automatic, uncommitted and yet computationally cheap. The algorithm is proposed based on a careful analysis of the roots of the spurious segments and on a morphological processing of the root

MMU GASPFA: A COTS multimodal biometric database

This paper describes the baseline corpus of a new multimodal biometric database, the MMU GASPFA (Gait-Speech-Face) database. The corpus in GASPFA is acquired using commercial off the shelf (COTS) equipment including digital video cameras, digital voice recorder, digital camera, Kinect camera and accelerometer equipped smart phones. The corpus consists of frontal face images from the digital camera, speech utterances recorded using the digital voice recorder, gait videos with their associated data recorded using both the digital video cameras and Kinect camera simultaneously as well as accelerometer readings from the smart phones. A total of 82 participants had their biometric data recorded. MMU GASPFA is able to support both multimodal biometric authentication as well as gait action recognition. This paper describes the acquisition setup and protocols used in MMU GASPFA, as well as the content of the corpus. Baseline results from a subset of the participants are presented for validation purposes.

Combined optimal wavelet filters with morphological watershed transform for the segmentation of dermoscopic skin lesions

In this paper, a technique is proposed to segment skin lesions from dermoscopic images through a combination of watershed transform and wavelet filters. In our technique, eight types of wavelet filters such as Daubechies and bi-orthogonal filters were applied before watershed transform. The resulting image was then classified into two classes: background and foreground. As watershed transform generated many spurious regions on the background, morphological post-processing was conducted. The post-processing split and merged spurious regions depending on a set of predefined criteria. As a result, a binary image was obtained and a boundary around the lesion was drawn. Next, the automatic boundary was compared with the manually delineated boundary by medical experts on 70 images with different types of skin lesions. We have obtained the highest accuracy of 94.61% using watershed transform with level 2 bi-orthogonal 3.3 wavelet filter. Thus, the proposed method has effectively achieved segmentation of the skin lesions, as shown in this paper.

Classification of human gait features with different apparel and walking speed

In this paper, we proposed a new approach for the classification of human gait features with different apparel and various walking speed. The approach consists of two parts: extraction of human gait features from enhanced human silhouette and classification of the extracted human gait features using fuzzy k-nearest neighbours (KNN). The joint angles together with the height, width and crotch height of the human silhouette are collected and used for gait analysis. The training and the testing sets are separable without overlapping. Both sets involve nine different apparel and three walking speed. From the experiment conducted, it can be observed that the proposed system is feasible as satisfactory results have been achieved.

Extraction and Classification of Human Gait Features

In this paper, a new approach is proposed for extracting human gait features from a walking human based on the silhouette images. The approach consists of six stages: clearing the background noise of image by morphological opening; measuring of the width and height of the human silhouette; dividing the enhanced human silhouette into six body segments based on anatomical knowledge; applying morphological skeleton to obtain the body skeleton; applying Hough transform to obtain the joint angles from the body segment skeletons; and measuring the distance between the bottom of right leg and left leg from the body segment skeletons. The angles of joints, step-size together with the height and width of the human silhouette are collected and used for gait analysis. The experimental results have demonstrated that the proposed system is feasible and achieved satisfactory results.

An improved image enhancement combining smoothing and sharpening

An effective method for contrast enhancement in an image was presented by Russo (2002), which was controlled by the trial-and-error tuning of one parameter. The same parameter was used for the entire image resulting in over blurring or sharpening, of features in the image. In this paper, we apply Russos algorithm on impulse noise and propose an efficient method for automatically obtaining the parameter value. Each pixel is adaptively assigned a different parameter value by evaluating the local features. Results of the proposed method are compared with those of Russos algorithm and of other methods for sharpening of image features. Experimental values indicate that the proposed method effectively tunes the operator yielding superior performance.

Lesion border detection in dermoscopy images using bilateral filter

This paper proposes an efficient way to effectively segment malignant melanoma in color dermoscopy images. A combination of methods are used in the proposed technique, including smoothing filters, PSNR, Spline, edge detection, morphological operations and segmentation. The pre-processing step eliminates noise, smoothes the image and employs the spline function to improve edge detection, while morphological operations are used to segment the lesion from image. Manual boundary selection is used as benchmark to test the accuracy of the automatic boundary selection by the proposed algorithm. The evaluation results show that the proposed method (Bilt-Sp) is able to achieve a good accuracy of 96.26%, supporting the effectiveness of the proposed method in automatically detecting skin lesions.

Adaptive parameter selection for improved fuzzy image enhancement

Previously, an effective parameter controlled method for noise reduction and sharpening of images corrupted by Gaussian noise was presented. However, the same parameter value was used for the entire image resulting in excess smoothing or sharpening of image features. In this paper, we adapted that method for impulse noise reduction and propose an efficient means to obtain the parameter value adoptively. After evaluating pixel local features through a fuzzy membership, each pixel location is assigned a different parameter value. Results of the proposed method are compared with the previous method and with results of other methods for sharpening of image features. Experimental results indicate that the proposed method has effectively assigned the parameter values, yielding superior performance.