Search

For the safe flight of drones, they must avoid the attacks of aggressive birds. These birds move very fast and must be detected far enough away. In recent years, deep learning has made it possible to detect small distant objects in RGB camera images. Since these methods are learning-based, they require a large amount of training images, but there are no publicly-available datasets for bird detection taken from drones. In this work, we propose a new dataset captured by a drone camera. Our dataset consists of 34,467 bird instances in 21,837 images that were captured in various locations and conditions. Our experimental results show that, even with the SOTA detection model, our dataset is sufficiently challenging. We also demonstrated that (1) several standard techniques for improving detection methods (e.g., data augmentation) are inappropriate for our challenging dataset, and (2) carefully-selected techniques can improve the detection performance.

An Around View Monitoring (AVM) system is widely used to allow a driver to watch the situation around a car. The AVM image is generated by image distortion correction and viewpoint transformation for images captured by wide view-angle cameras installed on the car. However, the AVM image is blurred due to these transformations. This blur impairs the visibility of the driver. While many deblurring methods based on CNN have been proposed, these general-purpose de-blurring methods are not designed for the AVM image. (1) Since the blur level in the AVM image is region-dependent, deblurring for the AVM should also be region-dependent. (2) Furthermore, while supervised deblurring methods require a pair of input-blurred and output-deblurred images, it is not easy to collect the deblurred AVM image. This paper proposes a method for generating the pairs of training images that cope with the aforementioned two problems. These training images are generated by the inverse transformation of the AVM image generation process. Experimental results show that our method can suppress blur on AVM images. We also confirmed that even a very shallow CNN with the inference time of 2.1ms has the same performance as the SoTA model.

This paper reviews the NTIRE2021 challenge on burst super-resolution. Given a RAW noisy burst as input, the task in the challenge was to generate a clean RGB image with 4 times higher resolution. The challenge contained two tracks; Track 1 evaluating on synthetically generated data, and Track 2 using real-world bursts from mobile camera. In the final testing phase, 6 teams submitted results using a diverse set of solutions. The top-performing methods set a new state-of-the-art for the burst super-resolution task., NTIRE 2021 Burst Super-Resolution challenge report

This paper presents a method for estimating object-scale proposals applied to super resolution (SR) for scale-optimized object detection. With the region-dependent scale proposals, we achieve scale-independent object detection. This object detection scheme consists of three functions; region-dependent scale proposals, SR, and object detection. While SR and object detection have been fused in deep end-to-end networks in previous works, region-dependent scale proposals are not provided or are performed independently of SR and object detection processes. The proposed region-dependent scale-proposal network is designed to explicitly estimate appropriate SR scales depending on the image region in accordance with scene contexts. Qualitative and quantitative experimental results show that our method can provide appropriate SR scales for improving detection accuracy. Our proposed method gains 2.7 points in AP with Centernet used as the base detector.

Lecture Notes in Computer Science, 12538, ISSN:0302-9743, ISSN:1611-3349, Computer Vision – ECCV 2020 Workshops Glasgow, UK, August 23–28, 2020, Proceedings, Part IV, ISBN:978-3-030-66822-8, ISBN:978-3-030-66823-5

Videos in the real-world contain various dynamics and motions that may look unnaturally discontinuous in time when the recorded frame rate is low. This paper reports the second AIM challenge on Video Temporal Super-Resolution (VTSR), a.k.a. frame interpolation, with a focus on the proposed solutions, results, and analysis. From low-frame-rate (15 fps) videos, the challenge participants are required to submit higher-frame-rate (30 and 60 fps) sequences by estimating temporally intermediate frames. To simulate realistic and challenging dynamics in the real-world, we employ the REDS_VTSR dataset derived from diverse videos captured in a hand-held camera for training and evaluation purposes. There have been 68 registered participants in the competition, and 5 teams (one withdrawn) have competed in the final testing phase. The winning team proposes the enhanced quadratic video interpolation method and achieves state-of-the-art on the VTSR task.

This paper reviews the NTIRE 2020 challenge on perceptual extreme super-resolution with focus on proposed solutions and results. The challenge task was to super-resolve an input image with a magnification factor 16 based on a set of prior examples of low and corresponding high resolution images. The goal is to obtain a network design capable to produce high resolution results with the best perceptual quality and similar to the ground truth. The track had 280 registered participants, and 19 teams submitted the final results. They gauge the state-of-the-art in single image super-resolution., Comment: CVPRW 2020

This paper reviews the AIM 2019 challenge on extreme image super-resolution, the problem of restoring of rich details in a low resolution image. Compared to previous, this challenge focuses on an extreme upscaling factor, ×16, and employs the novel DIVerse 8K resolution (DIV8K) dataset. This report focuses on the proposed solutions and final results. The challenge had 2 tracks. The goal in Track 1 was to generate a super-resolution result with high fidelity, using the conventional PSNR as the primary metric to evaluate different methods. Track 2 instead focused on generating visually more pleasant super-resolution results, evaluated using subjective opinions. The two tracks had 71 and 52 registered participants, respectively, and 9 teams competed in the final testing phase. This report gauges the experimental protocol and baselines for the extreme image super-resolution task.

This paper reviews the first NTIRE challenge on perceptual image enhancement with the focus on proposed solutions and results. The participating teams were solving a real-world photo enhancement problem, where the goal was to map low-quality photos from the iPhone 3GS device to the same photos captured with Canon 70D DSLR camera. The considered problem embraced a number of computer vision subtasks, such as image denoising, image resolution and sharpness enhancement, image color/contrast/exposure adjustment, etc. The target metric used in this challenge combined PSNR and SSIM scores with solutions' perceptual results measured in the user study. The proposed solutions significantly improved baseline results, defining the state-of-the-art for practical image enhancement.

This paper reviews the NTIRE 2019 challenge on real image denoising with focus on the proposed methods and their results. The challenge has two tracks for quantitatively evaluating image denoising performance in (1) the Bayer-pattern raw-RGB and (2) the standard RGB (sRGB) color spaces. The tracks had 216 and 220 registered participants, respectively. A total of 15 teams, proposing 17 methods, competed in the final phase of the challenge. The proposed methods by the 15 teams represent the current state-of-the-art performance in image denoising targeting real noisy images.

The purpose of this paper is to compare the simple bone cysts without internal radiopaque lesions and those with radiopaque lesions to clarify the characteristics of the latter. Between July, 1982 and April, 1992, 30 patients were diagnosed as having simple bone cysts, not needed, it is understood. Of the 30 patients, 7 females showed radiopaque lesions in the cysts. The mean age in the patients with simple bone cysts alone was 21.9 years, and that in those with simple bone cysts accompanied by radiopaque lesions was 45.9 years. The molar region was the most frequent site of both groups. Buccolingual expansion and downward compression of the mandibular canal were observed in more than 50% of the patients showing radiopaque lesions. Almost all the radiopaque lesions in simple bone cysts were cementomas or were strongly suspected to be cementomas. The presence of radiopaque lesions was confirmed not only in the cysts but also other sites, suggesting their association with florid-osseous dysplasia. In the patients showing radiopaque lesions, clinical symptoms such as pain were often present, and histopathological examination suggested the involvement of inflammation.

The diagnostic value of three-dimensional (3-D) CT imaging in various inflammatory conditions of the jaw bones was assessed in 6 patients and was compared together with high-resolution CT imaging and conventional radiographs. Though 3-D CT imaging provides no new information when viewing a series of axial high-resolution CT, 3-D CT could provide the enhanced perception of the changes of the bone surface topographically. With 3-D CT the defects on the bone surface and the spread of periosteal new bone formation were easier to interpret and more accessible, these could be overlooked even by high-resolution CT. 3-D CT images were variable up to the conditions of the threshold selection for 3-D reconstruction. Therefore, it was thought that it is very important to interpret the high-resolution CT used properly before the reconstruction procedure of 3-D CT imaging.

This is a retrospective study that reviewed 235 malignant cases of the oral and maxillo-facial region in Okayama University Hospital attached to Dental School between April 1982 and March 1988, of these it was found 6 cases of metastatic tumors to the mouth and jaws from a distant area. Malignant tumors of the lung provided the largest number of mouth and jaw metastasis in our series, the appearance of the lesions was as that of adenocarcinomas. Six patients were ranging in age from 42 to 82 years, and 3 were males and 3 were females. The most frequent symptoms observed were swelling and pain. Metastasis to the jaw bones were found predominantly in the mandible, and metastasis to the oral soft tissue occurred in two cases. Metastasis to the mandibular molar region radiographically showed central bone destruction, and metastasis to other regions showed peripheral bone destruction. Biopsy and radionuclide examinations were very useful for diagnosis of the metastatic tumors to the oral and maxillo-facial region.

Although osteomyelitis involving mandibular condylar process is scarcely mentioned in the past literature, we have encountered 11 cases of such kind of disease during past three years. These were consisted of 7 males and 4 females between the ages of 21 years and 77 years. The average age of them at the time of the initial diagnosis was 46.1 years. The most frequent cause was the extention of infection from pericoronitis of following extraction of the wisdom tooth. Prognosis was less favorable and it tended to run a pronounced chronic course of longer duration strongly resistant to the treatment. In radiographic analysis of bone changes of condylar process, there was a tendency to show that the more the infection extended upward, the more bone destruction became severe. It is emphasized that the disease of ramus and condylar process has the risk of being easily missed because of atypical signs and symptoms by the use of antibiotics and difficult radiographic interpretation.