一种基于多色彩通道动态阈值的舌苔舌质分离算法

舌苔舌质分离是实现舌诊图像自动分析重要的预处理工作,舌苔舌质分离后便于计算,消除相互的干扰,才能得到比较好的定量化结果.本文在结合中医依靠经验辨识舌苔舌质的基础上,借鉴以往的单通道定阈值舌苔舌质分离算法,提出使用多个色彩通道动态选取阈值,再根据先验的空间信息辅助判断,最终分离舌苔舌质的算法.通过与手工分离结果以及现有自动分离算法的比较,分析表明该算法在分离舌苔舌质的准确度以及处理速度上都有较好的效果.     

一种基于完全匹配和分词匹配的混合分词匹配算法

如何高效准确地进行酶的鉴别是代谢网络重构的难点和关键.由于基因标注文件中与ORF对应的蛋白质和酶信息都是采用自然语言描述的,重构中必须利用基因产物在酶数据库中寻找证据,从而进行与代谢相关的酶鉴别.针对目前通用的酶鉴别算法识别率低、效率低的问题,论文在对生物数据特性分析的基础上,提出了一种完全匹配和分词匹配相结合的混合分词匹配算法对酶信息进行鉴别,实验结果和分析表明,该算法具有更好的鉴别能力和更高的效率.     

基于直方图均衡化的伽马校正和K-means聚类的舌象苔质分离方法

目的舌苔、舌质分离对后续肿瘤患者舌象的客观化辨证具有重要的意义。常用的算法是基于颜色空间通道舌图像的K-means聚类算法。CIELAB颜色空间的a*通道舌图像相较于其他颜色空间通道的舌图像分割结果稳定,常用于后续的分割。对于部分舌图像,a*通道舌图像的舌苔、舌质虽然具有一定的区分度,但二者的区分度并不是十分明显,影响后续的分割结果。因此,本文提出一种基于直方图均衡化的伽马校正和K-means聚类的舌苔、舌质分离方法。方法采用200幅肿瘤患者的舌图像作为实验样本。首先将舌图像从RGB颜色空间转换到CIELAB颜色空间,对a*通道舌图像进行直方图均衡化增强以及伽马校正,然后利用K-means聚类方法对增强后的舌图像舌苔、舌质分离,得到直方图均衡化以及伽马校正后的a*通道舌图像和分割后的舌苔、舌质图像。为了验证算法的可行性,请5位专业中医医生对200例肿瘤患者的舌图像舌苔、舌质分割效果进行辨析。结果进行直方图均衡化以及伽马校正后的a*通道舌图像舌苔、舌质分割结果明显强于未经处理的a*通道舌图像分割结果。经辨析,分割合格率为97%。结论该方法可以很好地实现舌苔、舌质分离,具有一定的应用价值。     

中医舌象信息的数字化方法研究

提出一种基于图像处理的舌象检测与分析方法 ,以及研究传统中医舌诊中舌的特征的数字化途径。应用了 2 DGabor小波变换和色度信息较为精确地检测出舌体区域 ;提出了运用统计方法标定舌质和舌苔点以及确定其颜色的算法。舌苔的厚度是通过色度信息和 2 DGabor小波系数能量 (GWTE)进行量化。分析了在不同相位时 ,有无舌纹情况下GWTE呈现出的不同特性 ;根据这一特点用不变矩描述了舌区域的GWTE从而对舌纹的多少作出了定性的说明。通过一定的实验表明 ,本文提出的相关算法是行之有效的 ,并取得了较好的结果。     

图像分割技术在中医舌诊客观化研究中的应用

舌诊是中医四诊的主要内容,是辨证论治的主要依据。客观化研究对中医辨证规范化及中医临床、教学和科研手段的现代化具有重要意义。对舌诊客观化研究中涉及的图像预处理的重要内容——舌体分割提取和舌苔舌质同类区域划分——进行了深入研究,提出了相应算法,通过实验充分证明了算法具有很好的鲁棒性。这给进一步的自动特征提取提供了保障和重要信息。     

一种基于Gene Ontology注释信息的基因选择算法

基因选择算法是辅助生物学分析最重要的方法之一,但这类统计学算法受样本量相对基因数目过少的困扰.提出一种结合Gene Ontology(GO)注释信息的基因选择算法,用GO注释接近基因的方差的加权平均进行修正,增强小样本量下对总体的估计,进而寻找差异表达基因.将该算法与其他5种常见算法对比,以选择出的基因为特征构建分类器,以分类器的可靠性作为衡量算法的标准.3组芯片实验的结果表明,该算法在小样本情况下具有一定优势.亦有Pubmed文献证明,该算法可以鉴别出其他算法未曾发现的致病基因.该方法所建立起来的框架,是把生物学注释信息引入算法改进的一种有效尝试.     

基于混合遗传算法的心脏病决策支持系统研究

将遗传算法和 BP算法相结合 ,建立了一个基于混合遗传算法的心脏病决策支持系统来鉴别诊断五种常见心脏病 (冠心病 ,高血压性心脏病 ,风湿性心脏病 ,慢性肺原性心脏病和先天性心脏病 )。一个含有 35 2份心脏病的数据库用来构建和测试了该系统。实验结果表明 ,构建的系统对这五种心脏病均有较好的诊断识别率 ,系统的平均识别准确性达 90 .6 % ,各疾病的用户准确性和程序准确性均大于 85 .0 % ,表现出良好的心脏病的临床诊断决策支持能力     

脾动脉纤维肌肉发育不良1例并文献复习

目的 探讨脾动脉纤维肌肉发育不良(fibromuscular dysplasia,FMD)的临床病理特征、诊断和鉴别诊断.方法 对1例脾动脉FMD进行组织形态学、特殊染色检查,并复习相关文献.结果 该例临床呈急性脾脏动脉破裂出血,因失血过度而死亡.病理检查示脾动脉扭曲,管壁厚薄不均,内弹力层局灶消失,外膜与中膜部分分离.结论 脾动脉FMD罕见,早期临床表现缺乏特异性,主要依靠动脉造影和病理学来诊断.     

一种改进的中医舌苔厚度自动描述方法

目的舌诊客观化是中医现代化的重要内容之一,其中舌苔的量化描述是其重要指标。传统的舌苔薄厚分类只是定性地将舌苔分为薄厚两类,缺乏定量描述。本文提出一种灰度共生矩阵和小波纹理特征提取结合的方法。方法首先采用Daubechies2正交小波对舌图像进行一级小波分解,并求3个细节子图的平均值和方差作为特征,然后结合灰度共生矩阵在一级小波分解后的近似子图中提取三个方向的对比度、逆差距、能量和相关性作为补充特征,在此基础上训练支持向量机(support vector machine,SVM)分类器,对舌苔薄厚进行定性判断,同时建立支持向量回归(support vector regression,SVR)模型对舌苔薄厚定量分析。结果实验采取200例舌象样本,其中100例薄苔,100例厚苔,选取其中60幅薄苔和60幅厚苔样本进行分类器训练,其余80例样本作为测试。结论实验结果表明,与传统的小波纹理特征提取方法相比,本文提出的方法能够提升舌苔薄厚的分类效果,并在定量分析中取得较好的结果。     

基于GPU的超声成像算法的仿真

超声成像算法的仿真对于成像系统的研究和设计有重要的意义.我们在分析波束形成算法和图形处理器的并行性的基础上,以改进仿真速度为目标,提出了一种新的超声成像算法仿真方案.对基于GPU和基于CPU的仿真方案进行了对比实验,实验所得成像图像完全一致.通过实验证明新的超声成像算法仿真方案是可行的.     

Internal Kinematics of the Tongue Following Volume Reduction

This study was undertaken to determine the functional consequences following tongue volume reduction on tongue internal kinematics during mastication and neuromuscular stimulation in a pig model. Six ultrasonic‐crystals were implanted into the tongue body in a wedge‐shaped configuration which allows recording distance changes in the bilateral length (LENG) and posterior thickness (THICK), as well as anterior (AW), posterior dorsal (PDW), and ventral (PVW) widths in 12 Yucatan‐minipigs. Six animals received a uniform mid‐sagittal tongue volume reduction surgery (reduction), and the other six had identical incisions without tissue removal (sham). The initial‐distances among each crystal‐pairs were recorded before, and immediately after surgery to calculate the dimensional losses. Referring to the initial‐distance there were 3–66% and 1–4% tongue dimensional losses by the reduction and sham surgeries, respectively. The largest deformation in sham animals during mastication was in AW, significantly larger than LENG, PDW, PVW, and THICK (P < 0.01–0.001). In reduction animals, however, these deformational changes significantly diminished and enhanced in the anterior and posterior tongue, respectively (P < 0.05–0.001). In both groups, neuromuscular stimulation produced deformational ranges that were 2–4 times smaller than those occurred during chewing. Furthermore, reduction animals showed significantly decreased ranges of deformation in PVW, LENG, and THICK (P < 0.05–0.01). These results indicate that tongue volume reduction alters the tongue internal kinematics, and the dimensional losses in the anterior tongue caused by volume reduction can be compensated by increased deformations in the posterior tongue during mastication. This compensatory effect, however, diminishes during stimulation of the hypoglossal nerve and individual tongue muscles. Anat Rec, 291:886‐893, 2008. © 2008 Wiley‐Liss, Inc.     

中医舌色苔色识别中多分类支撑向量机算法比较

为了了解在不同样本数量情况下具有何种特点的算法比较适合舌色苔色的分类与识别,本文采用一对一SVM和它的改进算法IMP-BSVM、DAG-BSVM,以及M-ary分类SVM算法对舌色、苔色分类进行了比较。结果表明,当训练样本数较少时,DAG-BSVM能有较高的正确率和较快的训练和识别速度,因此选用DAG-BSVM;当训练样本数较大时,IMP-BSVM能得到最少的支撑向量和较高的正确率以及较快的识别速度,此时则选用IMP-BSVM。     

Simultaneous recording of ultrasound and polysomnography during natural sleep in patients with obstructive sleep apnea: a pilot study

Tongue base deformation may play a critical role in the phenomenon of obstructive sleep apnea, but polysomnography provides limited information regarding the effect of tongue motion during natural sleep. We reported on preliminary results of combining a novel ultrasound system and polysomnography for simultaneous recordings during natural sleep in volunteers and patients with obstructive sleep apnea. All participants underwent time‐synchronized polysomnography and submental transcutaneous ultrasound examinations. The wearable ultrasound device detected the air–mucosal interface of the tongue surface and automatically determined the maximum tongue base thickness in real time. All participants reported no sensation of heat, no sign of skin allergy, and an average of mild disturbance after the ultrasound recordings. In the individual patient with obstructive sleep apnea, we demonstrated a significant difference (P < 0.001) between the ultrasonic tongue base thickness measured during eupnea and that measured during snoring, hypopnea and apnea. The ultrasonic tongue base thickness increased and remained before the occurrence of obstructive apnea. On average, increased tongue base thicknesses of 2.5 (4.1%), 6.0 (9.8%) and 7.7 mm (12.5%) are associated with snoring, hypopnea and apnea, respectively. Our present data demonstrate that simultaneous examination of ultrasonic tongue base thickness and polysomnography is feasible for prolonged recording during natural sleep. The proposed method also enables the detection of significant differences in ultrasonic tongue base thickness between eupnea and obstructive respiratory events evaluated using polysomnography. This novel technique can be used to generate hypotheses for subsequent investigations of the underlying mechanisms and individualized combined therapy for obstructive sleep apnea. Clinical trial registration: This study has been registered at the Chinese Clinical Trial Registry website with the registration number of ChiCTR‐DDT‐13003313. The date of registration was 13 July 2013.     

Cesium Implant for Tongue Carcinoma with a Thickness of 1.5 cm or More: Cases Successfully Treated with a Modified Manchester System

Purpose

Deciding on treatment carcinoma of the tongue when the tumor has a thickness of 1.5 cm or more is difficult. Surgery often requires wide resection and re-construction, leading to considerable functional impairment. A cesium implant is an attractive option, but according to the Manchester System, a two plane implant is needed.

Materials and Methods

According to the textbook, a tumor is sandwiched between the needles, which are implanted at the edge of the tumor. This may cause an unnecessarily high dose to the outer surface of the tongue, which sometimes leads to a persistent ulcer. To avoid this complication, we invented a modified implantation method, and applied the method to five consecutive patients.

Results

With a minimum follow-up of 2 years, all primary tumors in 5 consecutive patients have been controlled. No complications occurred in soft tissue of the tongue or in the mandible.

Conclusion

Our modified Manchester System was feasible and effective for tumors that has a thickness of 1.5 cm or more.     

Distribution pattern of the human lingual nerve

The tongue is an intricate organ with many functions. Despite the knowledge of the presence of muscular and neural connections in the tongue, a detailed neuroanatomical depiction of the nerves' topography in the tongue has not been demonstrated. The topography, branching patterns and neuronal interconnections of the lingual nerve were studied in five postmortem human tongues. They were stained with Sihler's stain, a technique that renders most of the tongue tissue translucent while counterstaining nerves. The lingual nerve reaches the tongue posterolaterally. There are two main branches off of the main trunk: the medial branch sends 2-4 small branches to the medial part of the ventrolateral tongue and the lateral branch runs along the lateral tongue border and sends 3-4 large branches to the anterior tip of tongue. Each subdivision gives off 2-5 distal branches. Both medial and lateral branches have interconnections with the proximal part of the hypoglossal nerve. One of the unexpected discoveries in this study was the high density of nervous fibers in the lateral aspect of the tongue as compared to the midline region. The average diameter of the main trunk of the lingual nerve is 3.5 mm. The medial and lateral branches average 1 mm in diameter, the more distal subdivisions measure 0.5-0.75 mm, and the lingual-hypoglossal interconnections measure 0.125-0.250 mm. In summary, this study provides the first detailed depiction of the topography of the human lingual nerve and its branches in situ, confirmation of lingual-hypoglossal nerve connection, and the first depiction of the high density of lingual nerve innervation in the lateral tongue.     

满族5项舌运动类型的人类学研究

目的研究辽宁满族卷舌、叠舌、翻舌、尖舌、三叶舌5种舌运动类型的分布情况,为人类学群体遗传学积累资料。方法在辽宁抚顺清原县调查了401例(男197例,女204例),所得数据录入SSPS13.0软件包,进行频数分析、卡方检验等统计学处理。结果满族的尖舌、卷舌、叠舌、翻舌、三叶舌率分别是54.9%、52.4%、46.4%、19.5%、9.7%,除满族的叠舌和尖舌率男女性别间有明显差异外,其余舌运动类型无性别差异。结论与我国其他民族相比,辽宁满族卷舌、翻舌、尖舌、三叶舌出现率较低,叠舌出现率高。     

Template Guided Live Wire and Its Application on Automatic Extraction of Tongue in Digital Image

In this paper, we propose a novel automatic object extraction algorithm, named the Template Guided Live Wire, based on the popularly used livewire techniques. We discuss in details the novel method＇s applications on tongue extraction in digital images. With the guides of a given template curve which approximates the tongue＇ s shape, our method can finish the extraction of tongue without any human intervention. In the paper, we also discussed in details how the template guides the live wire, and why our method functions more effectively than other boundary based segmentation methods especially the snake algorithm. Experimental results on some tongue images are as well provided to show our method＇s better accuracy and robustness than the snake algorithm.