Text recognition in natural scene images is a challenging problem computer vision. Different than the optical character (OCR), text more complex due to variations size, colors, fonts, orientations, backgrounds, occlusion, illuminations and uneven lighting conditions. In this paper, we propose segmentation-free method based on deep convolutional recurrent neural network solve of cursive recognition, particularly focusing Urdu scenes. Compared non-cursive scripts, writing styles, several...

Detection of anomalies in video surveillance plays a key role ensuring the safety and security public spaces. The number cameras is growing, making it harder to monitor them manually. So, automated systems are needed. This change increases demand for that detect abnormal events or anomalies, such as road accidents, fighting, snatching, car fires, explosions real-time. These improve detection accuracy, minimize human error, make operations more efficient. In this study, we proposed Composite...

The extensive research in the field of multimodal biometrics by community and advent modern technology has compelled use real life applications. Biometric systems that are based on a single modality have many constraints like noise, less universality, intra class variations spoof attacks. On other hand, biometric gaining greater attention because their high accuracy, increased reliability enhanced security. This paper proposes develops Convolutional Neural Network (CNN) model for feature...

Reading text in natural scene images is an active research area the fields of computer vision and pattern recognition as detection, script identification are required. In this data article, a comprehensive dataset for Urdu detection presented analysed. To develop dataset, more than 2500 were captured using digital camera built-in mobile phone camera. Three separate datasets isolated character images, cropped word end-to-end spotting developed. The contain much larger number samples existing...

Multimodal biometrie systems have received greater attention from the research community since few decades due to various reasons such as: increase accuracy and efficiency as compared a single biometric system, decrease error rate or false acceptance improve security of data. In this research, preliminary idea has been discussed for fusion fingerprint online signature make data safe theft misuse. The proposed multimodal system will combine features both signature, which are most widely used...

This research study asserts the potential of ensemble learning classifiers over traditional classification approaches in field biometric recognition. Three traits: fingerprint, online signature and offline are considered to demonstrate improvement accuracies using classifiers. Data standardization techniques applied scale data a standard form. Histogram Oriented Gradients is extract features from then fed different such as Support Vector Machine (SVM), k-nearest neighbors algorithm (k-NN),...

Conventional solar cells are not economical and recently too expensive to the manufacturers for extensive-scale electricity generation. Cost efficiency is most vital factor in accomplishment of any technology. In order improve conversion efficiency, major research third generation photovoltaic (PV) directed toward retaining more sunlight using nanotechnology. Advancement nanotechnology cell via quantum dots (QDs) could reduce cost PV additionally enhance efficiency. Silicon (Si-QDs)...

Biometrics recognition plays a vital role in modern human and verification systems. An extensive latest research by the community has rendered field of biometrics inevitable for real-life applications. This study focuses on online signature recognition. The is performed to identify if an genuine or forged. A novel dataset, based 1000 signatures, been collected from 200 participants, wherein every participant provided 5 instances signature. Android-based mobile application was developed...

Handwritten signature for the identification and authentication of an individual has been widely used in biometric systems. Due to intra-class inter-class variabilities, verification become one most challenging problem technology. Furthermore, offline handwritten can be forged by skilled persons due its static nature. Therefore, this paper a deep learning-based method using convolutional neural network (CNN) online developed. Different values kernels such as 1×1, 3×3 5×5 are extract...

Now-a-days, in the field of machine learning data augmentation techniques are common use, especially with deep neural networks, where a large amount is required to train network. The effectiveness technique has been analyzed for many applications; however, it not separately multimodal biometrics. This research analyzes effects on single biometric and data. In this research, features from two modalities: fingerprint signature, have fused together at feature level. primary motivation fusing...

Text detection in natural images is a challenging problem due to variations text size, aspect ratio, alignment and background complexity. This paper proposes multiscale feature fusion convolutional neural network method detect cursive multi-language images. The proposed combines VGG-16 features at multi-scales multi-layers creates new map of shallow deep layers. On top map, vertical proposal generation used that generates fixed-size proposals. A recurrent layer implemented which takes the...