Plant diseases have been detrimental for the agriculture industry, as they cause substantial crop loss globally. To overcome this, IoT and AI-based smart agriculture solutions are being deployed for plant disease detection. However, a diverse range of crops and their diseases pose enormous challenges to these methods. Additionally, limited generalizability and the black-box nature of existing deep learning models, together with the scarcity of in-field datasets, are the main bottlenecks in developing efficient and acceptable solutions for large-scale applications. In the present work, a lightweight model 'ConViTX' is proposed for plant disease classification that demonstrates improved generalizability and explainability. The compact architecture of ConViTX uses a fusion of convolutional neural networks and vision transformers to simultaneously capture local and global features. Remarkably, ConViTX outperforms nine state-of-the-art deep learning methods on four publicly available datasets and a self-collected in-field maize dataset. Furthermore, the model demonstrates explainable prediction through Gradient Weighted Class Activation Maps and Locally Interpretable Model-Agnostic Evaluations. ConViTX attains 98.8% accuracy on the maize dataset and 61.42% on drone camera-captured raw images. With only 0.7 million parameters and 0.647 billion operations per second, the proposed model has the potential for deployment on resource-constrained precision agriculture setups.

A large number of emerging applications, such as autonomous navigation, space exploration, surveillance, military target detection, and remote sensing, use outdoor images to monitor various activities of interest. However, images acquired under unfavorable weather conditions usually suffer from atmospheric scattering due to environmental pollution causing color-shift and low-contrast images. Dehazing is an emerging research area in the computer vision domain that intends to restore the visibility of images by eliminating the latter types of degradation. Single image dehazing, on the other hand, is more challenging since it necessitates a precise assessment of atmospheric light and transmission map. This study aims to design a dual-channel deep neural network (DCD-Net) for estimating the transmission map, further utilized to compute atmospheric light. Finally, a dehazed image is generated using the obtained atmospheric light and the transmission map. The experimental results are compared qualitatively and quantitatively with eight existing dehazing approaches based on ten metrics on six publicly available standard datasets: Foggy Road Image DAtabase, HazeRD, REalistic Single Image DEhazing, NYU-Depth, O-HAZE, I-HAZE, a few natural hazy images, and underwater images. The DCD-Net outperforms conventional techniques, according to extensive studies. Moreover, a range of relative improvements of the proposed method over other approaches is calculated for better analysis of the results. A visual internet of things (VIoT) framework employing a PYNQ-Z2 board is presented in addition to the DCD-Net. It can be applied in real-time applications, particularly in the transportation and surveillance industries. The DCD-Net is suitable for image dehazing by virtue of its multilayered structure. The VIoT uses the DCD-Net for dehazing, while the PYNQ-Z2 board serves as the central processing unit. Note to Practitioners—This paper is motivated by the problems occurring due to haze. Haze reduces the visibility of a scene, causing major concerns in transportation and surveillance. Existing approaches have attempted to address this issue, albeit the methods are limited. As a result, this study proposes a new dual-channel CNN model with two modules, where the first module calculates fine details of the image and the second module estimates the transmission map. Furthermore, both features are combined to produce a more reliable transmission map. The training process highly influences the resulting output of the network. Therefore, an algorithm explaining the training instructions for the practitioners is given in the appendix. The obtained transmission map is further used to estimate atmospheric light. The images are then dehazed using atmospheric light and transmission maps. In addition, we have designed a framework for image dehazing using VIoT with a PYNQ-Z2 board. Experimental results suggest that this approach gives expected results, yet, there is one limitation. This method requires a haze image and a corresponding transmission map, which is not always possible. Therefore, we will attempt to design a semi-supervised learning approach in the future.

Visibility issues in intelligent transportation systems are exacerbated by bad weather conditions such as fog and haze. It has been observed from recent studies that major road accidents have occurred in the world due to low visibility and inclement weather conditions. Single image dehazing attempts to restore a haze-free image from an unconstrained hazy image. We proposed a dehazing method by cascading two models utilizing a novel parameter-adaptive dual-channel modified simplified pulse coupled neural network (PA-DC-MSPCNN). The first model uses a new color channel for removing haze from images. The second model is the improved brightness preserving model (I-GIHE), which retains the brightness of the image while improving the gradient strength. To integrate the results from these two models and provide a pleasing haze-free image, a PA-DC-MSPCNN-based fusion is used. Furthermore, the proposed approach is deployed on a Xilinx Zynq SoC by exploiting the recently released PYNQ platform. The dehazing system runs on a PYNQ-Z2 all-programmable SoC platform, where it will input the camera feed through the FPGA unit and carry out the dehazing algorithm in the ARM core. This configuration has allowed reaching real-time processing speed for image dehazing. The results of dehazing are analyzed using both synthetic and real-world hazy images. Synthetic hazy images are acquired from the O-HAZE, I-HAZE, SOTS, and FRIDA datasets, while real-world hazy images are taken from the RailSem19, E-TUVD dataset, and the internet. For evaluation, twelve cutting-edge approaches are chosen. The proposed method is also analyzed on underwater and low-light images. Extensive experiments indicate that the proposed method outperforms state-of-the-art methods of qualitative and quantitative performances.

Polyps are very common abnormalities in human gastrointestinal regions. Their early diagnosis may help in reducing the risk of colorectal cancer. Vision-based computer-aided diagnostic systems automatically identify polyp regions to assist surgeons in their removal. Due to their varying shape, color, size, texture, and unclear boundaries, polyp segmentation in images is a challenging problem. Existing deep learning segmentation models mostly rely on convolutional neural networks that have certain limitations in learning the diversity in visual patterns at different spatial locations. Further, they fail to capture inter-feature dependencies. Vision transformer models have also been deployed for polyp segmentation due to their powerful global feature extraction capabilities. But they too are supplemented by convolution layers for learning contextual local information. In the present paper, a polyp segmentation model CoInNet is proposed with a novel feature extraction mechanism that leverages the strengths of convolution and involution operations and learns to highlight polyp regions in images by considering the relationship between different feature maps through a statistical feature attention unit. To further aid the network in learning polyp boundaries, an anomaly boundary approximation module is introduced that uses recursively fed feature fusion to refine segmentation results. It is indeed remarkable that even tiny-sized polyps with only 0.01% of an image area can be precisely segmented by CoInNet. It is crucial for clinical applications, as small polyps can be easily overlooked even in the manual examination due to the voluminous size of wireless capsule endoscopy videos. CoInNet outperforms thirteen state-of-the-art methods on five benchmark polyp segmentation datasets.

With the advent of deep learning, the research on facial expression recognition (FER) has received a lot of interest. Different deep convolutional neural network (DCNN) architectures have been developed for real-time and efficient FER. One of the challenges in FER is obtaining trustworthy features that are strongly associated with facial expression changes. Furthermore, traditional DCNNs for FER problems have two significant issues: insufficient training data, which leads to overfitting, and intra-class facial appearance variations. FLEPNet, a texture-based feature-level ensemble parallel network for FER, is proposed in this study and proved to solve the aforementioned problems. Our parallel network FLEPNet uses multi-scale convolutional and multi-scale residual block-based DCNN as building blocks. First, we consider modified homomorphic filtering to normalize the illumination effectively, which minimizes the intra-class difference. The deep networks are then protected against having insufficient training data by using texture analysis on face expression images to identify multiple attributes. Four texture features are extracted and combined with the image's original characteristics. Finally, the integrated features retrieved by two networks are used to classify seven facial expressions. Experimental results reveal that the proposed technique achieves an average accuracy of 0.9914, 0.9894, 0.9796, 0.8756, and 0.8072 on Japanese Female Facial Expressions, Extended CohnKanade, Karolinska Directed Emotional Faces, Real-world Affective Face Database, and Facial Expression Recognition 2013 databases, respectively. Moreover, experimental outcomes depict significant reliability when compared to competing approaches.

Pulse coupled neural network (PCNN) is widely used in image fusion framework due to its global coupling and pulse synchronization of neurons. However, its manual setting of parameters and inability to process multiple images affect the fusion performance. In this letter, a novel weighted parameter adaptive dual channel PCNN (WPADCPCNN) based medical fusion method is proposed in non-subsampled shearlet transform domain to fuse the magnetic resonance imaging and single-photon emission computed tomography images of AIDS dementia complex and Alzheimer's disease patients. The parameters of the proposed WPADCPCNN model are estimated from its inputs using fractal dimension. The high-pass sub-bands are fused using the WPADCPCNN model whereas the low-pass sub-bands are merged using a new weighted multi-scale morphological gradients based rule. Experimental results demonstrate that the proposed method outperforms some of the state-of-the-art methods in terms of both visual quality and objective assessment.

Abstract This paper presents a new hybrid fusion framework based on thermal and visible face images. Fusion of information is done here in two phases, first at the pixel level and then at the decision level. For the pixel level fusion process, à‐trous wavelet transform is applied on both the thermal and visible face images. In decision level fusion, 34 region classifiers, each concentrating on a specified region of the face image, are tested individually for their ability to identify a person from the face image. The region classifiers, which contribute significantly in recognizing the face image, are considered for decision level fusion using majority voting. All experiments have been conducted on the UGC‐JU face database and IRIS benchmark face database. The maximum recognition rate is about 97.22% for both the databases whereas decisions of 17 region classifiers among 34 are considered. Experimental results and comparative study show that the proposed fusion method provides a framework for recognition of face images in uncontrolled environments such as variations in illumination conditions, pose, and facial expressions.

An illumination and expression invariant face recognition method based on uniform local binary patterns (uLBP) and Legendre moments is proposed in this work. The proposed method exploits uLBP texture features and Legendre moments to make a feature representation with enhanced discriminating power. The input images are preprocessed to extract the face region and normalized. From normalized image, uLBP codes are extracted to obtain texture image which overcomes the effect of monotonic temperature changes. Legendre moments are computed from this texture image to get the required feature vector. Legendre moments conserve the spatial structure information of the texture image. The resultant feature vector is classified using k-nearest neighbor classifier with [Formula: see text] norm. To evaluate the proposed method, experiments are performed on IRIS and NVIE databases. The proposed method is tested on both visible and infrared images under different illumination and expression variations and performance is compared with recently published methods in terms of recognition rate, recall, length of feature vector, and computational time. The proposed method gives better recognition rates and outperforms other recent face recognition methods.

A new image fusion algorithm based on the visible and thermal images for face recognition is presented in this paper. The new fusion algorithm derives the benefit from both the modalities images. The proposed fusion process is the weighted sum of thermal and visible face information with two weighting factors [Formula: see text] and [Formula: see text], respectively. The weighting factors are calculated using a directed search algorithm automatically. The proposed fusion framework is evaluated through extensive experiments using UGC-JU face database. Experiments are of three fold. Firstly, individual modalities images are used separately for human face recognition. Secondly, fused face images using the proposed method are used for recognition purpose. The highest level of accuracy achieved by using the proposed method is about 98.42%. Lastly, the three existing fusion methods are applied on the same face database for comparison with the results of the proposed method. All the results demonstrate significant performance improvements in recognition over individual modalities and some of the existing fusion approaches, suggesting that fusion is a viable approach that deserves further study and consideration.

This paper presents a robust approach for recognition of thermal face images based on decision level fusion of 34 different region classifiers. The region classifiers concentrate on local variations. They use singular value decomposition (SVD) for feature extraction. Fusion of decisions of the region classifier is done by using majority voting technique. The algorithm is tolerant against false exclusion of thermal information produced by the presence of inconsistent distribution of temperature statistics which generally make the identification process difficult. The algorithm is extensively evaluated on UGC-JU thermal face database, and Terravic facial infrared database and the recognition performance are found to be 95.83% and 100%, respectively. A comparative study has also been made with the existing works in the literature.