The accurate segmentation of breast ultrasound images is an important precondition for the lesion determination. The existing segmentation approaches embrace massive parameters, sluggish inference speed, and huge memory consumption. To tackle this problem, we propose T ²KD Attention U-Net (dual-Teacher Knowledge Distillation Attention U-Net), a lightweight semantic segmentation method combined double-path joint distillation in breast ultrasound images. Primarily, we designed two teacher models to learn the fine-grained features from each class of images according to different feature representation and semantic information of benign and malignant breast lesions. Then we leveraged the joint distillation to train a lightweight student model. Finally, we constructed a novel weight balance loss to focus on the semantic feature of small objection, solving the unbalance problem of tumor and background. Specifically, the extensive experiments conducted on Dataset BUSI and Dataset B demonstrated that the T ²KD Attention U-Net outperformed various knowledge distillation counterparts. Concretely, the accuracy, recall, precision, Dice, and mIoU of proposed method were 95.26%, 86.23%, 85.09%, 83.59%and 77.78% on Dataset BUSI, respectively. And these performance indexes were 97.95%, 92.80%, 88.33%, 88.40% and 82.42% on Dataset B, respectively. Compared with other models, the performance of this model was significantly improved. Meanwhile, compared with the teacher model, the number, size, and complexity of student model were significantly reduced (2.2×10 ⁶ vs. 106.1×10 ⁶, 8.4 MB vs. 414 MB, 16.59 GFLOPs vs. 205.98 GFLOPs, respectively). Indeedy, the proposed model guarantees the performances while greatly decreasing the amount of computation, which provides a new method for the deployment of clinical medical scenarios.
Humans ; Breast Neoplasms/diagnostic imaging* ; Female ; Ultrasonography, Mammary/methods* ; Image Processing, Computer-Assisted/methods* ; Algorithms ; Neural Networks, Computer ; Breast/diagnostic imaging*

Objective To compare the clinical values of impedance method,optical method and microscopy when used to detect platelet abnormal results.Methods Platelet re-examinations by optical method and microscopy were carried out in case of low confidence degree in platelet test with impedance method by XE-2100 automatic hematology analyzer,and then the results by the three methods were compared.Results Most of the low-confidence-degree results by impedance method could be corrected by re-examination by optical method,and the remained had to turn to microscopy due to unsatisfied requirements of the instrument.Conclusion Optical method has to be involved to correct the platelet abnormal results by XE2100 automatic hematology analyzer,and microscopy should be applied in case optical method doesn't work.The three methods gains advantages and disadvantages of themselves,and can be supplementaries for one another.

Objective To explore the effect of adipose-derived stem cells (ADSCs) on inflammatory factors in rats with lipopolysaccharide (LPS)-induced acute lung injury (ALI) and the possible mechanism of anti-inflammatory. Methods Seventy male Sprague-Dawley (SD) rats were randomly divided into normal control group (n = 10), LPS model group (n = 30), and ADSCs intervention group (n = 30) by random number table. ALI model was reproduced by intraperitoneal injection of 8 mg/kg LPS, and the rats in ADSCs intervention group received tail vein injection of 300 μL ADSCs 30 minutes after the model reproduction, the samples of normal control group were harvested immediately without any intervention, and the specimens in remained two groups were taken at 6, 24, 72 hours respectively. Arterial partial pressure of oxygen (PaO2) and lactate level in femoral artery were determined. Enzyme-linked immunosorbent assay (ELISA) was used to detect the serum myeloperoxidase (MPO) and interleukin-10 (IL-10) in the blood of left ventricle. Lung wet/dry weight (W/D) ratio was detected by thoracotomy, and the pathological changes of lung tissue were observed under an optical microscope. Western Blot was used to detect the protein expression of nuclear factor-κB (NF-κB) in lung tissue of rats. Results Compared with the normal control group, the damage degree of lung tissue of LPS model group was significantly heavier from 6 hours, and lung W/D ratio, blood lactate, MPO, IL-10 and expression level of NF-κB in lung tissue were significantly increased respectively, while PaO2 was decreased significantly. Compared with LPS model group, the damage degree of lung tissue of ADSCs intervention group was significantly reduced from 6 hours, and lung W/D ratio, blood lactate, MPO, and NF-κB expression in lung tissue were significantly decreased, while PaO2 was increased significantly, and it became normal at 72 hours [lung W/D ratio: 5.33±0.29 vs. 5.77±0.42 at 6 hours, 5.14±0.46 vs. 5.43±0.38 at 72 hours; blood lactate (mmol/L): 3.6±1.0 vs. 5.7±1.1 at 6 hours, 3.1±1.0 vs. 3.8±1.2 at 72 hours; blood MPO (μg/L): 1.50±0.90 vs. 2.70±1.85 at 6 hours, 0.46±0.30 vs. 0.71±0.22 at 72 hours; NF-κB (gray value): 0.40±0.11 vs. 0.50±0.09 at 6 hours, 0.24±0.03 vs. 0.33±0.06; PaO2 (mmHg, 1 mmHg = 0.133 kPa): 78.0±4.1 vs. 74.5±3.2 at 6 hours, 89.3±9.4 vs. 81.9±3.4 at 72 hours; all P < 0.05]. The IL-10 level was significantly higher than that of LPS model group only at 24 hours (ng/L: 27.75±15.49 vs. 17.52±6.56, P < 0.05). Conclusion ADSCs can effectively relieve the inflammatory response of ALI induced by LPS, probably by inhibiting the expressions of NF-κB and blocking the release of inflammatory cytokines.