Night-time image enhancement (NIE) aims at boosting the intensity of low-light regions while suppressing noises or light effects in night-time images, and numerous efforts have been made for this ...task. However, few explorations focus on the quality evaluation issue of enhanced night-time images (ENTIs), and how to fairly compare the performance of different NIE algorithms remains a challenging problem. In this paper, we firstly construct a new Real-world Night-Time Image Enhancement Quality Assessment (i.e., RNTIEQA) dataset that includes two typical types of night-time scenes (i.e., extremely low light and uneven light scenes), and carry out human subjective studies to compare the quality of ENTIs obtained by a set of representative NIE algorithms. Afterwards, a new objective ranking method that comprehensively considering image intrinsic and impairment attributes is proposed for automatically predicting the quality of ENTIs. Experimental results on our RNTIEQA dataset demonstrate that the proposed method outperforms the off-the-shelf competitors. Our dataset and code will be released at https://github.com/Leilei-Huang-work/RNTIEQA-dataset .
Information security has become a key issue of public concern recently. In order to radically resist the decryption and analysis in the field of image information hiding and significantly improve the ...security of the secret information, a novel coverless information hiding approach based on deep learning is proposed in this paper. Deep learning can select the appropriate carrier according to requirements to achieve real-time image data hiding and the high-level semantic features extracted by CNN are more accurate than the low-level features. This method does not need to employ the designated image for embedding the secret data but transfer a set of real-time stego-images which share one or several visually similar blocks with the given secret image. In this approach, a group of real-time images searched online are segmented according to specific requirements. Then, the DenseNet is used to extract the high-level semantic features of each similar block. At the same time, a robust hash sequence with feature sequence, DC and location is generated by DCT. The inverted index structure based on the hash sequence is constructed to attain real-time image matching efficiently. At the sending end, the stego-images are matched and sent through feature retrieval. At the receiving end, the secret image can be recovered by extracting similar blocks through the received stego-images and stitching the image blocks according to the location information. Experimental results demonstrate that the proposed method without any modification traces provides better robustness and has higher retrieval accuracy and capacity when compared with some existing coverless image information hiding.
•Objective assessment of night-time images by proposing a blind night-time image quality assessment metric using exposure and gradient magnitude feature maps.•Unsupervised learning of local quality ...aware features using sparse coding which are more discriminative than hand-crafted ones.•Composition of local sparse representations by a spatial pyramid maximization pooling method to provide effective global quality-aware descriptions.•constructing an ensemble of weak learners through least squares gradient boosting for regression of subjective quality scores with high performance.
Capturing Night-Time Images (NTIs) with high-quality is quite challenging for consumer photography and several practical applications. Thus, addressing the quality assessment of night-time images is urgently needed. Since there is no available reference image for such images, Night-Time image Quality Assessment (NTQA) should be done blindly. Although Blind natural Image Quality Assessment (BIQA) has attracted a great deal of attention for a long time, very little work has been done in the field of NTQA. Due to the capturing conditions, NTIs suffer from various complex authentic distortions that make it a challenging field of research. Therefore, previous BIQA methods, do not provide sufficient correlation with subjective scores in the case of NTIs and special methods of NTQA should be developed. In this paper we conduct an unsupervised feature learning method for blind quality assessment of night-time images. The features are the sparse representation over the data-adaptive dictionaries learned on the image exposure and gradient magnitude maps. Having these features, an ensemble regression model trained using least squares gradient boosting scheme predicts high correlated objective scores on the standard datasets.
The MobileNetV3 is specially designed for mobile devices with limited memory and computing power. To reduce the network parameters and improve the network inference speed, a new lightweight network ...is proposed based on MobileNetV3. Firstly, to reduce the computation of residual blocks, a partial residual structure is designed by dividing the input feature maps into two parts. The designed partial residual structure is used to replace the residual block in MobileNetV3. Secondly, a dual-path feature extraction structure is designed to further reduce the computation of MobileNetV3. Different convolution kernel sizes are used in the two paths to extract feature maps with different sizes. Besides, a transition layer is also designed for fusing features to reduce the influence of the new structure on accuracy. The CIFAR-100 dataset and Image Net dataset are used to test the performance of the proposed partial residual structure. The ResNet based on the proposed partial residual structure has smaller parameters and FLOPs than the original ResNet. The performance of improved MobileNetV3 is tested on CIFAR-10, CIFAR-100 and ImageNet image classification task dataset. Comparing MobileNetV3, GhostNet and MobileNetV2, the improved MobileNetV3 has smaller parameters and FLOPs. Besides, the improved MobileNetV3 is also tested on CPU and Raspberry Pi. It is faster than other networks Keywords: MobileNetV3, Real-time image classification, Lightweight network, Deep convolutional neural network, residual structure
Image downscaling is arguably the most frequently used image processing tool. We present an algorithm based on convolutional filters where input pixels contribute more to the output image the more ...their color deviates from their local neighborhood, which preserves visually important details. In a user study we verify that users prefer our results over related work. Our efficient GPU implementation works in real-time when downscaling images from 24 M to 70 k pixels. Further, we demonstrate empirically that our method can be successfully applied to videos.
«What would Sophia Loren do?» is a question that Nancy Kulik, daughter of Italian immigrants, asked herself many times in her life. Even at a young age, Loren has served her not only as a compass for ...important choices, but also to create an emotional connection with her Italian heritage. Through its personal approach, the documentary «What would Sophia Loren do?» makes us understand how Kulik is thus representative of a nostalgic migrant’s view of the Italian-American diaspora. This short cinematic homage gives the opportunity to reflect on Sophia Loren’s cult of the star from the perspective of the “transatlantic-gaze” (Carolan McDonald, 2014), gender and time-image (Deleuze, 1991), and not least, “italianità” and “americanità” against the backdrop of the Italian-American migration experience.
This study introduces a lightweight visual segmentation model named TinySegformer, specifically designed for agricultural pest detection, aiming to address edge computing issues in real-world ...scenarios. The innovation of TinySegformer lies in its effective combination of Transformers and neural networks, enhancing its ability to handle semantic segmentation tasks and significantly improving image segmentation performance. Additionally, through the use of sparse attention mechanisms and quantization techniques, it adopts a lightweight architecture, enabling the model to adapt to the computational and storage limitations of edge devices. This study evaluates TinySegformer on multiple datasets, revealing that whether on public datasets or our self-collected datasets, TinySegformer outperforms current mainstream visual segmentation models such as DeepLab, SegNet, UNet, PSPNet (Pyramid Scene Parsing Network), FCN (Fully Convolutional Networks), etc. In terms of key performance indicators like precision, recall, accuracy, mIoU, Dice coefficient, and FPS, TinySegformer shows outstanding results, reaching 0.92, 0.90, 0.93, 0.85, 0.91, and 65 respectively, and achieves real-time image processing at 32.7 frames per second on edge devices. Furthermore, the study elaborately discusses the process of deploying TinySegformer onto NVIDIA Jetson devices, and successfully adapts the model to resource-constrained devices through network pruning and quantization techniques. In conclusion, with its efficiency, accuracy, and lightweight characteristics, the TinySegformer model provides a robust and practical solution for agricultural pest detection, offering new insights and directions for future research in the field of agricultural pest monitoring.
•A dataset with 102 pest types enables precise pest detection and actionable insights in real-world applications.•TinySegformer’s combination of Transformer and neural network technologies enhances image segmentation accuracy.•Its lightweight architecture is tailored for edge computing, ensuring precise pest detection in complex environments.•The model’s performance is thoroughly assessed using diverse metrics like accuracy, recall, precision, mIoU, and FPS.•Adapted for mobile devices through network pruning and quantization, it achieves real-time image processing at 32.7 FPS.
•Quality-aware features to characterize the night-time image quality were designed.•Semantic features of the image were extracted by deep neural network.•A blind night-time image quality metric was ...proposed.
High-quality night-time imaging is crucial to video surveillance, automatic drive and consumer electronics. However, different from day-time imaging, night-time imaging suffers from some disadvantages, such as low light, uneven illumination, difficult focusing, etc., which raises a great concern to the night-time imaging quality. Accordingly, a practical night-time image quality evaluation method is very promising to control and improve the night-time imaging system. Toward this end, in this paper, we propose a blind image quality assessment (BIQA) method to quantify the night-time image quality. Specifically, in the proposed method, we measure the night-time image quality by investigating the fundamental image properties, which are highly relevant to the image quality, such as the brightness, saturation, sharpness, noiseness, contrast and the semantics. Specific features are designed to characterize the image properties properly. Then we employ the support vector regression (SVR) method to infer the image quality with the extracted quality-aware features. The proposed BIQA method for night-time images is thoroughly evaluated on a representative night-time image database. Experimental results demonstrate that the proposed BIQA method for night-time images achieves superior prediction performance to other state-of-the-art BIQA methods.