Search

At present, deep learning has achieved great success in the field of computer vision, but small object detection is still a challenging problem in the field of object detection. Aiming at the problems of low resolution of small objects, blurred images, and less information carried, one object detector that introduces residual learning and multi-scale feature enhancement is proposed. Firstly, an enhanced feature mapping block based on residual learning is introduced into the backbone network. Through channel averaging and normalization, the model more focuses on the object area instead of the background, and it provides additional semantics information for the effective feature layer while taking into account the detection speed. Then the feature map increases the receptive field of the effective feature map through feature fusion block sensitive to context information, and fuses the shallow feature layer and the deep feature layer used for prediction to improve the detection performance at low resolution. Finally, a dual attention block is used to suppress background noise, and key features are embedded in attention. While preserving spatial information, it strengthens the information association between channels, thereby enhancing the expressive ability of features. In order to better detect small objects, the number of a priori boxes for shallow feature mapping is also adjusted. Experimental results show that on the dataset of PASCAL VOC2007, the detection accuracy (mAP) of the algorithm for 300×300 input scale is 79.9%, which is 2.7 percentage points higher than that of SSD, and the detection accuracy of small objects bird, bottle, chair, and plant is improved 5.1 percentage points, 7.5 percentage points, 3.9 percentage points, 7.2 percentage points，respectively. The detection accuracy (mAP) on the OAP self-made aerial dataset is 82.7%.

For the observation of a picture, people may instinctly pay more attention to the eye-catching objects in the picture. Usually such objects tend to occupy a larger proportion in the picture, which leads to small targets being ignored. Because the area where the small target is located is often a weak detection area, and the features that can be extracted in the process of extracting features by the detector are few and are easily lost in the process of feature information transmission after the feature is extracted, the effect of small target detection is not good. Therefore, on the basis of the single-order detector, this paper adds a cross-channel interaction mechanism to ensure the integrity of the information between layers, adopts target enhancement of training samples and designs a general loss function. Apart from this, this paper improves the sample weighting on the basis of the loss function to predict weight of samples. The mAP of this paper framework UWN (unified weighted network) on the VOC public dataset is 81.2% and the mAP on the self-made small target aerial photography dataset is 82.3%. Compared with the FSSD algorithm, some speed is sacrificed, and the accuracy is greatly improved.

In order to solve the problems of deadlock, slow convergence, easy to get into local optimum and uneven path when ant colony algorithm is applied to the path planning of mobile robots, a smooth path planning method combining improved A* ant colony algorithm and the rolling window method is proposed. Firstly, the improved A* algorithm is used to initialize the ant colony pheromone to solve the problem of low ant colony efficiency. Then, the state transition probability function is improved to consider the feasible path “activity” and the end position in the function to avoid deadlock phenomenon. At the same time, based on the mechanism of inequality principle, the pheromone of ant colony is updated to avoid falling into the local optimal path and accelerate the convergence speed of the algorithm. Secondly, based on the global path planning, local real-time path planning is carried out by integ-rating the rolling window method and the dynamic obstacle avoidance strategy. Finally, Bessel curve is used to process the smoothness of the planned path, so that the smoothed path is closer to the actual motion path. In order to ensure the best performance of the algorithm, genetic algorithm with elite strategy is used to optimize the parameters of the algorithm. Three sets of experimental results show that the proposed algorithm is effective in the presence of simple or complex, static or dynamic obstacles.

In order to solve the problems of deadlock, slow convergence, easy to get into local optimum and uneven path when ant colony algorithm is applied to the path planning of mobile robots, a smooth path planning method combining improved A* ant colony algorithm and the rolling window method is proposed. Firstly, the improved A* algorithm is used to initialize the ant colony pheromone to solve the problem of low ant colony efficiency. Then, the state transition probability function is improved to consider the feasible path"activity"and the end position in the function to avoid deadlock phenomenon. At the same time, based on the mechanism of inequality principle, the pheromone of ant colony is updated to avoid falling into the local optimal path and accelerate the convergence speed of the algorithm. Secondly, based on the global path planning, local real-time path planning is carried out by integrating the rolling window method and the dynamic obstacle avoidance strategy. Finally, Bessel curve is used to process the smoothness of the planned path, so that the smoothed path is closer to the actual motion path. In order to ensure the best performance of the algorithm, genetic algorithm with elite strategy is used to optimize the parameters of the algorithm. Three sets of experimental results show that the proposed algorithm is effective in the presence of simple or complex, static or dynamic obstacles. [ABSTRACT FROM AUTHOR]

Objective: To observe the effect of early enteral nutrition (EEN) on nutritional indicators and clinical outcomes in patients with severe heart failure undergoing mechanical ventilation., Methods: Thirty-four patients with severe heart failure (grade III-IV of cardiac function) and pulmonary infections undergoing mechanical ventilation admitted to intensive care unit (ICU) of Second Affiliated Hospital of Zhejiang University of Traditional Chinese Medicine from May 2017 to April 2018 were enrolled. They were randomly divided into EEN group and delayed enteral nutrition (DEN) group. Both groups were given routine treatment, including mechanical ventilation, improvement of cardiac function, anti-infection, protection of vital organ function, regulation of blood sugar and adjustment of electrolyte and acid-base balance. The patients in EEN group received enteral nutrition (EN) within 48 hours after ICU admission, and in DEN group, EN was started after the patients had spent the early stage of stress and had stable vital signs (48 hours after ICU admission). The changes in serum B-type natriuretic peptide (BNP), C-reactive protein (CRP), nutritional indicators and liver function indicators at ICU admission and 7 days after treatment were compared between the two groups. The time needed for patients to reach EN target, duration of mechanical ventilation, length of ICU stay, and the 28-day mortality were recorded, and complications were observed., Results: There was no significant difference in serum BNP, CRP, nutritional indicators or liver function indicators at ICU admission between the two groups. After treatment for 7 days, BNP and CRP in both groups were decreased significantly as compared with those at ICU admission [BNP (ng/L): 592.1±370.9 vs. 2 517.7±1 163.4 in EEN group, 621.9±418.8 vs. 2 251.5±1 006.8 in DEN group; CRP (mg/L): 46.0±19.6 vs. 59.8±22.5 in EEN group, 40.5±18.8 vs. 61.2±24.6 in DEN group, all P < 0.05], pre-albumin (PA) and transferrin (TF) were significantly increased [PA (g/L): 0.18±0.05 vs. 0.15±0.06 in EEN group, 0.17±0.04 vs. 0.12±0.06 in DEN group; TF (g/L): 1.6±0.4 vs. 1.5±0.4 in EEN group, 1.7±0.5 vs. 1.4±0.5 in DEN group, all P < 0.05]. However, there was no significant difference in the above indicators after treatment between the two groups (all P > 0.05). There was no significant change in liver function after treatment in both groups. The EN treatment was successfully completed in both groups. Some patients developed abdominal distension and diarrhea in varying degrees, which were alleviated by slowing down the infusion rate, supplemented by gastrointestinal motility drugs and intestinal flora adjustment drugs. The time needed to reach EN target in EEN group was significantly earlier than that in DEN group (hours: 42.4±10.2 vs. 53.8±17.1, P < 0.05), the duration of mechanical ventilation (days: 14.2±8.7 vs. 13.4±7.9), the length of ICU stay (days: 17.8±6.7 vs. 18.3±5.6) and 28-day mortality [5.9% (1/17) vs. 11.8% (2/17)] showed no significant difference as compared with those in DEN group (all P > 0.05), and it did not increase the incidence of aspiration pneumonia [23.5% (4/17) vs. 17.7% (3/17), P > 0.05]., Conclusions: EEN could help to achieve nutritional goals as soon as possible, improve the nutritional status of the body, and provide conditions and basis for further treatment of severe heart failure patients.

Objective: To investigate the clinical features of ocular syndromes in patients with neurobrucellosis., Method: This is a retrospective series case study. The clinical data of 5 patients with neurobrucellosis, who were treated in Department of Neurology, Beijing Tongren Hospital, Captical Medical Uinversity, from May 2009 to January 2015, were collected. Their epidemiological information, clinical manifestation, laboratory and radiologic examination, therapy and prognosis were analyzed., Results: Among the five patients, there were 3 males and 2 females, and their ages ranged from 20 to 67 years. The median age was 25 years. All patients had ever exposed to sheep. 1 patient lived close to a slaughterhouse and 2 patients lived in epidemic areas of brucellosis. 5 patients presented with binocular vision loss, 8 eyes with fundus edema, 3 patients with ophthalmoplegia;4 patients with fever,4 patients with headache,3 patients with neck stiffness,1 patient with movement and sensation disorders. Cerebrospinal fluid (CSF) pressure elevated, white cell number and protein increased with glucose reduction were detected respectively in 3 cases. While, CSF chloride decreased in 2 cases. Serum agglutination test for brucella was positive in 5 patients. Serum brucella culture was positive in 1 patient and CSF brucella culture was positive in 1 patient. Brain magnetic resonance imaging (MRI) showed that the optic nerve was involved in 3 patients,the meninges were involved in 1 patient and the brain white matter was involved in 1 patient. The combination of rifamycin, tetracycline, ceftriaxone sodium or quinolone were given to all patients and showed appreciated effects., Conclusions: The clinical features of ocular syndromes are atypical in patients with neurobrucellosis. Vision loss and ophthalmoplegia are more common to be seen. Neurobrucellosis should be considered when patients with ocular signs and other system symptoms without a definite diagnosis.