Mohammed Nazrul Islam

Super-resolution enhancement technique for low resolution video

Super-resolution enhancement is the technique of increasing spatial resolution of an image from the image itself or from several noisy images of the same scene but with additional information such as motion. This paper presents a kernel regression approach to reconstruct a high resolution image from several low resolution video frames. In many applications including cell phone and webcam the imaging sensors capture low resolution images due to low cost sensors or physical limitation of the hardware. So, the proposed resolution enhancement technique could be used as an inexpensive software alternative. The performance of the proposed algorithm is evaluated with several grayscale and color video streams and found successful when compared to other state of the art techniques.

Enhanced class associative generalized fringe-adjusted joint transform correlation for multiple target detection

A modified class associative fringe-adjusted filter-based technique is proposed for multiple target detection, where the number of processing steps remains fixed irrespective of the number of objects in the class. An enhanced version of generalized fringe-adjusted filters is developed for correlation improvement. Again, a shifted phase encoding technique is employed for generation of a single correlation peak per target object. The effectiveness of the proposed technique is verified by computer simulation for binary as well as gray-level images both with and without noise.

Reduction of four-wave mixing effect in an optical wavelength-division multiplexed system by utilizing different channel spacing and chromatic dispersion schemes

The performance of an optical wavelength-division multi- plexed system is analyzed taking the four-wave mixing (FWM) effect into account. The efficiency of FWM generation is observed to greatly reduce if the spacing between channels is made unequal. However, the scheme requires an increased bandwidth, which can further be reduced by utiliz- ing another scheme called repeated unequal channel spacing. Also the FWM efficiency is found to reduce in the presence of chromatic disper- sion. Use of nonuniform chromatic dispersion along the fiber length, which reduces the FWM effect while maintaining a low value of disper- sion is investigated. Various other dispersion management mechanisms are also considered in addition to different channel spacing schemes to achieve a comparative study as to which scheme gives the best solution to the FWM problem.

Optical encryption and multiplexing of personal identification information using orthogonal code

A new technique is proposed for optical encryption and multiplexing of binary characters and images used for personal identification information. Different binary images can first be encrypted using orthogonal code and then multiplexed together in the spatial domain. The resulting encrypted single image provides security as well as makes efficient use of storage and/or transmission capacity. The image can finally be decrypted and the individual input images can be decoded employing the same orthogonal code set. Because of the orthogonal nature of the code used, the encryption and decryption processes do not deteriorate the quality of the images employed. Also, the proposed technique involves a very simple architecture, as it does not require any mathematical transformation.

Optical security system employing quadrature multiplexing

A novel information security system is developed, employing quadrature multiplexing, to be implemented using optoelectronic devices. Two information signals are encrypted using the same code but two orthogonal functions, and then they are multiplexed together in the same domain. With orthogonal functions having zero cross-correlation between them, the encrypted information remains unaffected. Each encryption and multiplexing process can accommodate two information signals for a single code and hence for a single storage cell or transmission channel. The same process can be performed in multiple steps to increase the multiplexing capability of the system. The proposed security system can enhance the storage capacity and/or maximize the utilization of the available transmission bandwidth. For decryption, the composite encoded signal is correlated using the appropriate code and function. Computer simulation results show that the proposed security system is capable of dealing with both binary and gray-level images with the encrypted images remaining secure.

Optical pattern recognition using multiple phase- shifted-reference fringe-adjusted joint transform correlation

We propose a novel optical pattern recognition system using multiple phase-shifted-reference fringe-adjusted joint transform correlation (MRFJTC) techniques. The MRFJTC algorithm can efficiently detect an object of interest in the input scene by producing a highly distinctive correlation peak while rejecting any and all nontarget objects in a complex background. The simple architecture of the proposed system can simultaneously recognize multiple targets of the reference class when present. The recognition performance is fast, automatic, and invariant to noise and distortions.

Optical cryptographic system employing multiple reference-based joint transform correlation technique

An optical joint transform correlation-based cryptographic system is a used to feed multiple phase-shifted encryption keys into four parallel channels along with a to-be-encrypted signal in the form of an image. The resulting joint power spectra (JPS) signals are phase-shifted and then combined to yield a modified JPS signal. Inverse Fourier transformation of the modified JPS signal yields the secured encrypted image. For decryption purpose, the received encrypted signal is first Fourier transformed and multiplied by the encryption key used in encryption. The derived signal is then inverse Fourier transformed to generate the output signal. The proposed system offers a nonlinear encryption without the involvement of any complex mathematical operation on the encryption key otherwise required in similar encryption techniques and is invariant to noise. Computer simulation results are presented to show the effectiveness of the proposed scheme with binary, as well as gray images in both noise-free and noisy environment.

Distortion-invariant multiple target detection using class-associative joint transform correlation

A distortion-invariant class-associative pattern recognition technique is proposed, where a class of objects may be defined as a group of objects with similarity and dissimilarity among them. The fractional power fringe-adjusted joint transform correlation technique as well as the synthetic discriminant function concept has been effectively utilized to achieve the distortion-invariant detection of multiple dissimilar targets simultaneously present in the input scene. Simulation results prove that the proposed scheme is an effective tool for the detection of multiple dissimilar targets in both binary and gray-level input scenes corrupted by distortion and noise.

Single image super-resolution in frequency domain

This paper presents a neighborhood dependent components based feature learning (NDCFL) for regression analysis in single image super-resolution. Given a low resolution input, the method uses directional Fourier phase feature components to adaptively learn the regression kernel based on local covariance to estimate the high resolution image. Although this formulation resembles other regression and covariance based methods, our method uses image features to learn the local covariance from geometric similarity between low resolution image and its high resolution counterpart. For each patch in the neighborhood, we estimate four directional variances to adapt the interpolated pixels. Experimental results show that the proposed algorithm performs better than other state of the art techniques especially at higher resolution scales.

Anomaly based vessel detection in visible and infrared images

Detection of small vessels is a challenging task for navy, coast guard and port authority for security purposes. Vessel identification is more complex as compared to other object detection because of its variability in shapes, features and orientations. Current methods for vessel detection are primarily based on segmentation techniques which are not as efficient and also require different algorithms for visible and infrared images. In this paper, a new vessel detection technique is proposed employing anomaly detection. The input intensity image is first converted to feature space using difference of Gaussian filters. Then a detector filter in the form of Mahalanobis distance is applied to the feature points to detect anomalies whose characteristics are different from their surroundings. Anomalies are detected as bright spots in both visible and infrared image. The larger the gray value of the pixels the more anomalous they are to be. The detector output is then post-processed and a binary image is constructed where the boat edges with strong variance relative to the background are identified along with few outliers from the background. The resultant image is then clustered to identify the location of the vessel. The main contribution in this paper is developing an algorithm which can reliably detect small vessels in visible and infrared images. The proposed method is investigated using real-life vessel images and found to perform excellent in both visible and infrared images with the same system parameters.