Polyp Detection from Colorectum Images by Using Attentive YOLOv5

Background: High-quality colonoscopy is essential to prevent the occurrence of colorectal cancers. The data of colonoscopy are mainly stored in the form of images. Therefore, artificial intelligence-assisted colonoscopy based on medical images is not only a research hotspot, but also one of the effective auxiliary means to improve the detection rate of adenomas. This research has become the focus of medical institutions and scientific research departments and has important clinical and scientific research value. Methods: In this paper, we propose a YOLOv5 model based on a self-attention mechanism for polyp target detection. This method uses the idea of regression, using the entire image as the input of the network and directly returning the target frame of this position in multiple positions of the image. In the feature extraction process, an attention mechanism is added to enhance the contribution of information-rich feature channels and weaken the interference of useless channels; Results: The experimental results show that the method can accurately identify polyp images, especially for the small polyps and the polyps with inconspicuous contrasts, and the detection speed is greatly improved compared with the comparison algorithm. Conclusions: This study will be of great help in reducing the missed diagnosis of clinicians during endoscopy and treatment, and it is also of great significance to the development of clinicians’ clinical work.

Recurrent Video Deblurring with Blur-Invariant Motion Estimation and Pixel Volumes

For the success of video deblurring, it is essential to utilize information from neighboring frames. Most state-of-the-art video deblurring methods adopt motion compensation between video frames to aggregate information from multiple frames that can help deblur a target frame. However, the motion compensation methods adopted by previous deblurring methods are not blur-invariant, and consequently, their accuracy is limited for blurry frames with different blur amounts. To alleviate this problem, we propose two novel approaches to deblur videos by effectively aggregating information from multiple video frames. First, we present blur-invariant motion estimation learning to improve motion estimation accuracy between blurry frames. Second, for motion compensation, instead of aligning frames by warping with estimated motions, we use a pixel volume that contains candidate sharp pixels to resolve motion estimation errors. We combine these two processes to propose an effective recurrent video deblurring network that fully exploits deblurred previous frames. Experiments show that our method achieves the state-of-the-art performance both quantitatively and qualitatively compared to recent methods that use deep learning.

DEVELOPMENT OF A FRANGIBLE DESIGN OF SMALL FIXED-WING UNMANNED AERIAL SYSTEM

Existing studies show that small fixed-wing unmanned aircraft systems’ (FWUASs) mid-air collisions with aircraft can cause substantial damage. Upon a 250 knots impact, a ~1.8 kg “tractor” configuration of FW-UAS can perforate aircraft skin, thereby damaging the internal structures such as ribs, frames, etc., posing severe threat to manned air fleet. Significant damage is primarily caused by FW-UAS’s heavy and rigid components such as motor, battery, and payload especially due to their roughly in-line arrangement and proximity with one another. In this work, a modified FW-UAS finite element (FE) model was developed that included a “pusher” engine (i.e., motor in the aft of the forward fuselage) configuration to reduce the impact severity during airborne collisions. A polymeric foam nosecone was attached to the front of the FW-UAS FE model to dissipate impact energy. To assess its energy absorbing capacity, a comparative study with expanded polypropylene (EPP), polyurethane (PUR), and polystyrene (IMPAXX700) foams was performed. Conical and semi-spherical nosecone configurations were studied as part of this research. A series of LS-Dyna impact simulations were performed with the pusher configuration of FW-UAS impacting a 1.59 mm thick aluminum 2024-T3 flat plate sandwiched between a rigid target frame. In addition, a frangible design of the FW-UAS, in which the payload is diverged from the in-line collision trajectory of battery and motor upon impact, was implemented and assessed. Force generated during the initial stage of impact is leveraged through lightweight and friable structural links to diverge the payload to avoid impact along the single axis as of the battery and motor. Damage severity is evaluated through target plate tear, and velocity of payload during impact, it being the major damage causing component.

Deep Learning-Based Multitarget Motion Shadow Rejection and Accurate Tracking for Sports Video

The effect is tested in various specific scenes of sports videos to complete the multitarget motion multitarget tracking detection application applicable to various specific scenes within sports videos. In this paper, deep neural networks are applied to sports video multitarget motion shadow suppression and accurate tracking to improve tracking performance. After the target frame selection is determined, the tracker uses an optical flow method to estimate the limits of the target sports video multitarget motion based on the sports video multitarget motion of the target object between frames. The detector first scans each sports video image frame one by one, observing the previously discovered and learned image frame subregions one by one until the current moment that is highly like the target to be tracked. The preprocessed remote sensing images are converted into grayscale images, the histogram is normalized, and the appropriate height threshold is selected in combination with the regional growth function to realize the rejection of sports video multitarget motion shadow and establish the sports video multitarget network model. The distance and direction of the precise target displacement are determined by frequency-domain vectors and null domain vectors, and the target action judgment mechanism is formed by decision learning. Finally, comparing with other shadow rejection and precision tracking algorithms, the proposed algorithm achieves greater advantages in terms of accuracy and time consumption.

Immature Apple Detection Method Based on Improved Yolov3

The identification of immature apples is a key technical link to realize automatic real-time monitoring of orchards, expert decision-making, and realization of orchard output prediction. In the orchard scene, the reflection caused by light and the color of immature apples are highly similar to the leaves, especially the obscuration and overlap of fruits by leaves and branches, which brings great challenges to the detection of immature apples. This paper proposes an improved YOLOv3 detection method for immature apples in the orchard scene. Use CSPDarknet53 as the backbone network of the model, introduce the CIOU target frame regression mechanism, and combine with the Mosaic algorithm to improve the detection accuracy. For the data set with severely occluded fruits, the F1 and mAP of the immature apple recognition model proposed in this article are 0.652 and 0.675, respectively. The inference speed for a single 416×416 picture is 12 ms, the detection speed can reach 83 frames/s on 1080ti, and the inference speed is 8.6 ms. Therefore, for the severely occluded immature apple data set, the method proposed in this article has a significant detection effect, and provides a feasible solution for the automation and mechanization of the apple industry.

Immature Apple Detection Method Based on Improved Yolov3

The identification of immature apples is a key technical link to realize automatic real-time monitoring of orchards, expert decision-making, and realization of orchard output prediction. In the orchard scene, the reflection caused by light and the color of immature apples are highly similar to the leaves, especially the obscuration and overlap of fruits by leaves and branches, which brings great challenges to the detection of immature apples. This paper proposes an improved YOLOv3 detection method for immature apples in the orchard scene. Use CSPDarknet53 as the backbone network of the model, introduce the CIOU target frame regression mechanism, and combine with the Mosaic algorithm to improve the detection accuracy. For the data set with severely occluded fruits, the F1 and mAP of the immature apple recognition model proposed in this article are 0.652 and 0.675, respectively. The inference speed for a single 416×416 picture is 12 ms, the detection speed can reach 83 frames/s on 1080ti, and the inference speed is 8.6 ms. Therefore, for the severely occluded immature apple data set, the method proposed in this article has a significant detection effect, and provides a feasible solution for the automation and mechanization of the apple industry.

Prophylactic effect of the aqueous extract of Pimpinella anisum on the behavior of Wistar rats exposed to mercury

The purpose of this study was to evaluate the prophylactic effect of Pimpinella anisum (green anis) on neurobehavioral status following mercury chloride intoxication during the developmental period. For this purpose, rats exposed to 100 mg/L of HgCl2 during the gestation and lactation period. A group of rats was treated with the anis extract for 15 days before becoming intoxicated with mercury. In contrast, one group was orally administered aqueous anis extract for 15 days after intoxication. The forced swimming test, the open field test and the Morris pool respectively recorded an increase in immobility time, a decrease in the number of cross-cells (p <0.001), (p <0.05) and an increase in latency (p <0.01), (p <0.001), (p <0.001) and decreased time spent in the target frame during the probe test (p <0.01) and increased latency in the visible test (p <0.01) in HgCl2 - exposed rats compared to control rats. However, preventive and curative aniseed-based treatment reduced the rate of depression, increased locomotor activity and improved learning performance. In conclusion, the aqueous extract of Pimpinella anisum could have a corrective effect on some neurological disorders caused by mercury.

Evaluating the effectiveness of a multivariate autoregressive model in predicting the time course of dynamic facial behaviours

Faces convey critical information about people, such as cues to their identity and emotional state. In the real world, facial behaviours evolve dynamically and encapsulate a range of biological motion signals. Furthermore, behavioural and neuroimaging studies have demonstrated that human observers are sensitive to this temporal information. The presence of systematic temporal changes in the face implies the possibility of predicting the evolution of dynamic facial behaviours. We video recorded subjects delivering positive or negative phrases, and used a PCA-based active appearance model to capture critical dimensions of facial variation over time. We applied multivariate autoregressive models to predict PCA scores of future frames from the frames immediately preceding them, up to a lag of 200ms prior to the target frame. These models did successfully predict future frames, but they did not benefit from extending the temporal support, suggesting they relied primarily on image similarity between consecutive frames. We next used hidden Markov models to segment videos into shorter sequences comprising more consistent facial behaviours. The Markov models successfully extracted distinct facial basis states, however segmenting the data by state did not yield any predictive benefit to autoregressive models fit within those states. We conclude that autoregressive models have only limited predictive power in the context of facial expression analysis.

A Semantic Web-based Representation of Human-logical Inference for Solving Bongard Problems

Bongard Problems (BPs) are a set of 100 visual puzzles introduced by M. M. Bongard in the mid-1960s. BPs have been established as benchmark puzzles for understanding the human context-based learning abilities to solve ill- posed problems. The puzzle requires the logical explanation as the answer to distinct two classes of figures from redundant options, which can be obtained by a thinking process to alternatively change the target frame (hierarchical level of analogy) of thinking from a wide range concept networks as D. R. Hofstadter suggested. Some minor research results to solve a limited set of BPs have reported based a single architecture accompanied with probabilistic approaches; however the central problem on BP's difficulties is the requirement of flexible changes of the target frame, therefore non-hierarchical cluster analyses does not provide the essential solution and hierarchical probabilistic models needs to include unnecessary levels for learning from the beginning to prevent a prompt decision making. We hypothesized that logical reasoning process with limited numbers of meta-data descriptions realizes the sophisticated and prompt decision-making and the performance is validated by using BPs. In this study, a semantic web-based hierarchical model to solve BPs was proposed as the minimum and transparent system to mimic human-logical inference process in solving of BPs by using the Description Logic (DL) with assertions on concepts (TBox) and individuals (ABox). Our results demonstrated that the proposed model not only provided individual solutions as a BP solver, but also proved the correctness of Hofstadter's idea as the flexible frame with concept networks for BPs in our actual implementation, which no one has ever achieved. This fact will open the new horizon for theories for designing of logical reasoning systems especially for critical judgments and serious decision-making as expert humans do in a transparent and descriptive way of why they judged in that manner.

Temporal Adaptive Alignment Network for Deep Video Inpainting

Video inpainting aims to synthesize visually pleasant and temporally consistent content in missing regions of video. Due to a variety of motions across different frames, it is highly challenging to utilize effective temporal information to recover videos. Existing deep learning based methods usually estimate optical flow to align frames and thereby exploit useful information between frames. However, these methods tend to generate artifacts once the estimated optical flow is inaccurate. To alleviate above problem, we propose a novel end-to-end Temporal Adaptive Alignment Network(TAAN) for video inpainting. The TAAN aligns reference frames with target frame via implicit motion estimation at a feature level and then reconstruct target frame by taking the aggregated aligned reference frame features as input. In the proposed network, a Temporal Adaptive Alignment (TAA) module based on deformable convolutions is designed to perform temporal alignment in a local, dense and adaptive manner. Both quantitative and qualitative evaluation results show that our method significantly outperforms existing deep learning based methods.