The bag‐of‐visual words (BOVWs) have been recognised as an effective mean of representing images for image classification. However, its reliance on a visual codebook developed using handcrafted image feature extraction algorithms and vector quantisation via k ‐means clustering often results in significant computational overhead, and poor classification accuracies. Therefore, this study presents an adaptive BOVW modelling, in which image feature extraction is achieved using deep feature learning and the amount of computation required for the development of visual codebook is minimised using a batch implementation of particle swarm optimisation. The proposed method is tested using Caltech‐101 image dataset, and the results confirm the suitability of the proposed method in improving the categorisation performance while reducing the computational load.

The massive amount of digital content generated daily in the modern world has created the need for an image retrieval system built on image analysis via image processing and machine learning, therefore this study explains the role of machine learning in bridging the semantic gap in content‐based image retrieval, proposes an automatic image annotation framework, in which training images are obtained from social media, and semantic indexing is achieved using a combination of supervised and unsupervised machine learning. Furthermore, the study also highlights the need for continuous vocabulary improvement for optimum system performance and recommends hardware implementation of machine learning algorithms to ensure high overall speed of image retrieval systems.

Detection of motor intention from brain signals combined with robot assistive technologies has potential to be used as an effective rehabilitation process for post-stroke patients. The work conducted on the deployment of AMADEO hand rehabilitation robotic device and Electroencephalogram based Brain Computer Interference (EEG-BCI) system to explore the technical feasibility of the approach in hand motor recovery of post-stroke patients is presented. Two different protocols consisting of simple visual cues and a 2D interactive game are presented to healthy subjects when performing hand movement. The motor intent signals produced during each protocol are detected using Support Vector Machine (SVM) algorithm. Moreover, the signals produced by different single electrodes are analyzed to identify the electrode making the highest contribution to the intent signal and the performance of SVM with respect to each protocol. Overall, an average True Positive Rate (TPR) of 71.72% and True Negative Rate (TNR) of 63.33% for visual cue protocol and an average TPR of 88.56% and TNR of 70.81% for game protocol are obtained.

Assessing the frailty of older people quantitatively is critical to prevent potential accidents and to ensure their well-being. The older people with high frailty score are at the risk of fall, which increases the rate of hospitalization and reducing the number of independent activities carried out. The conventional clinical tools used for frailty assessment are subjective, qualitative and are prone to human error. The balance assessment, activity of daily living (ADL) and gait analysis are practiced as clinical and quantitative tools for risk of fall and frailty assessments. An objective approach to classify the frailty levels using ADL is proposed. The "pick up an object from floor" as an ADL is deployed to differentiate the signal patterns obtained through inertial measurement unit (IMU) for frail and non-frail subjects. The data from single inertial unit mounted on pelvis is analyzed. The experimental work is carried out on three groups of healthy/control, frail and non-frail subjects. The various signal attributes are used to classify the frailty quantitatively using IMU data and machine learning methods. The results demonstrate that frail subjects have clear irregularities in their signal trajectories. Using the proposed algorithm two classes of frailty (non-frail and frail) are identified objectively. The study demonstrates the potential of deploying IMU for advanced classification of frailty levels in older people.

This work presents a novel method for image similarity measure, where a hand-drawn rough black and white sketch is compared with an existing data base of full color images (art works and photographs). The proposed system creates ambient intelligence in terms of the evaluation of nonprecise, easy to input sketched information. The system can then provide the user with options of either retrieving similar images in the database or ranking the quality of the sketch against a given standard, i.e., the original image model. Alternatively, the inherent pattern-matching capability of the system can be utilized to allow detection of distortion in any given real time-image sequences in vision-driven ambient intelligence applications. The proposed method can cope with images containing several complex objects in an inhomogeneous background. Two abstract images are obtained using strong edges of the model image and the morphologically thinned outline of the sketched image. The angular-spatial distribution of pixels in the abstract images is then employed to extract new compact and effective features using the Fourier transform. The extracted features are rotation and scale invariant and robust against translation. Experimental results from seven different approaches confirm the efficacy of the proposed method in both the retrieval performance and the time required for feature extraction and search.

This paper presents a novel approach for sketch-based image retrieval based on low-level features. It enables the measuring of the similarity among full color multi-component images within a database (models) and simple black and white user sketched queries. It needs no cost intensive image segmentation. Strong edges of the model image and morphologically thinned version of the query image are used for image abstraction. Angular-radial decomposition of pixels in the abstract images is used to extract new compact and affine invariant features. Comparative results, employing an art database (ArT BANK), show significant improvement in average normalized modified retrieval rank (ANMRR) using the proposed features.

This work presents a novel fast method for line segment extraction, based on a chain code representation of edge maps. It has a parallel nature and can be employed on parallel machines. In the first phase it breaks the macro chains into several micro chains after applying shifting, smoothing and differentiating. The micro chains are then approximated by straight line segments. In the second phase, based on the length and the error criteria, the line segments are grouped into much longer lines. The experimental results show a significant improvement in the number of line segments extracted while their accumulative length keeps high.

A new approach for document image analysis and verification is presented. The approach utilizes connected component analysis and geometric properties of labelled regions for region of interest extraction. Document images containing Persian/Arabic text combined with English text, headlines, ruling lines, trade mark and cursive signature are used as a test data. Persian/Arabic signature extraction is investigated as a case study. The proposed method uses special characteristics of such signatures for extraction and verification procedures. A set of efficient, invariant and compact features is extracted utilizing spatial partitioning of the signature region. Comparative results exhibit high extraction and verification rates

An approach is described for design and development of robotic system controllers. The technique, referred to as multiprocessing control, can provide more expansibility, flexibility, and portability in a robotic system than conventional multiprocessor methods. The control algorithms for a Mitsubishi robot are implemented as a multiprocessing system on a VME-bus-based computer running the PDOS operating system. The control software is written entirely in a real-time C and consists of concurrent tasks synchronized with events, semaphores, and messages. A proposal for implementing the technique as a hardware-independent software on the Inmos Transputer is given.>