Springer Science and Business Media LLC
Công bố khoa học tiêu biểu
* Dữ liệu chỉ mang tính chất tham khảo
Sắp xếp:
Multiscale self-quotient filtering for an improved unsupervised retinal blood vessels characterisation
Springer Science and Business Media LLC - - 2017
Digital images often suffer from contrast variability and non-uniform illumination, which seriously affect the evaluation of biomarkers such as the arteriolar to venular ratio. This biomarker provides valuable information about many pathological conditions such as diabetes, hypertension etc. Hence, in order to efficiently estimate the biomarkers, correct classification of retinal vessels extracted from digital images, into arterioles and venules is an important research problem. This paper presents an unsupervised retinal vessel classification approach which utilises the multiscale self-quotient filtering, to pre-process the input image before extracting the discriminating features. Thereafter the squared-loss mutual information clustering method is used for the unsupervised classification of retinal vessels. The proposed vessel classification method was evaluated on the publicly available DRIVE and INSPIRE-AVR databases. The proposed unclassified framework resulted in 93.2 and 88.9% classification rate in zone B for the DRIVE and the INSPIRE-AVR dataset respectively. The proposed method outperformed other tested methods available in the literature. Retinal vessel classification, in an unsupervised setting is a challenging task. The present framework provided high classification rate and therefore holds a great potential to aid computer aided diagnosis and biomarker research.
A wearable system for adaptation to left–right reversed audition tested in combination with magnetoencephalography
Springer Science and Business Media LLC - Tập 7 - Trang 205-213 - 2017
Exposure of humans to unusual spaces is effective to observe the adaptive strategy for an environment. Though adaptation to such spaces has been typically tested with vision, little has been examined about adaptation to left–right reversed audition, partially due to the apparatus for adaptation. Thus, it is unclear if the adaptive effects reach early auditory processing. Here, we constructed a left–right reversed stereophonic system using only wearable devices and asked two participants to wear it for 4 weeks. Every week, the magnetoencephalographic responses were measured under the selective reaction time task, where they immediately distinguished between sounds delivered to either the left or the right ear with the index finger on the compatible or incompatible side. The constructed system showed high performance in sound localization and achieved gradual reduction of a feeling of strangeness. The N1m intensities for the response-compatible sounds tended to be larger than those for the response-incompatible sounds until the third week but decreased on the fourth week, which correlated with the initially shorter and longer reaction times for the compatible and incompatible conditions, respectively. In the second week, disruption of the auditory-motor connectivity was observed with the largest N1m intensities and the longest reaction times, irrespective of compatibility. In conclusion, we successfully produced a high-quality space of left–right reversed audition using our system. The results suggest that a 4-week exposure to the reversed audition causes optimization of the auditory-motor coordination according to the new rule, which eventually results in the modulation of early auditory processing.
Recent trends in two-photon auto-fluorescence lifetime imaging (2P-FLIM) and its biomedical applications
Springer Science and Business Media LLC - Tập 9 - Trang 293-310 - 2019
Two photon fluorescence microscopy and the numerous technical advances to it have served as valuable tools in biomedical research. The fluorophores (exogenous or endogenous) absorb light and emit lower energy photons than the absorption energy and the emission (fluorescence) signal is measured using a fluorescence decay graph. Additionally, high spatial resolution images can be acquired in two photon fluorescence lifetime imaging (2P-FLIM) with improved penetration depth which helps in detection of fluorescence signal in vivo. 2P-FLIM is a non-invasive imaging technique in order to visualize cellular metabolic, by tracking intrinsic fluorophores present in it, such as nicotinamide adenine dinucleotide, flavin adenine dinucleotide and tryptophan etc. 2P-FLIM of these molecules enable the visualization of metabolic alterations, non-invasively. This comprehensive review discusses the numerous applications of 2P-FLIM towards cancer, neuro-degenerative, infectious diseases, and wound healing.
Plasmon based super resolution imaging for single molecular detection: Breaking the diffraction limit
Springer Science and Business Media LLC - Tập 4 - Trang 231-238 - 2014
Developments of super-resolution imaging techniques have considerable interests to detect and image tiny molecular events under a diffraction limit. Stimulated emission and depletion (STED) microscopy, photo-activated light microscopy (PALM) and structured illumination microscopy (SIM) are representative successful novel imaging techniques. Recently, surface plasmons (SP) based super resolution imaging techniques which can achieve super resolution with no deviation from conventional microscopic schematics have been actively investigated. In this paper, we explain the principle of SP phenomena which can apply for bioimaging, and introduce localized SP based super resolution imaging techniques to increase lateral and axial resolution below the diffraction limits. Three different novel techniques based on field localization are introduced to increase lateral resolution. Also, additional three imaging techniques based on extraordinary transmission and Förster resonance energy transfer are introduced to increase axial resolution. Consequently, we explore a future direction of SP based imaging researches for 3D spatiotemporal super resolution microscopy.
Development of FPGA-based coincidence units with veto function
Springer Science and Business Media LLC - Tập 1 - Trang 27-31 - 2011
A coincidence detector is an essential unit to acquire distribution of radioisotopes in positron emission tomography (PET) systems. The coincidence pairs rapidly increase as the number of detector modules increase. To make this system compact, flexible, and conveniently controllable, a field programmable gated array (FPGA)-based coincidence detector was developed. The unit targeted PET systems have eight detector modules. The coincidence detector module needs veto function for controlling the data transfer rate of the output signal because the output is connected to the band-limited ADC and data transfer unit. The coincidence detector module has eight inputs connected to the detector modules, one coincidence output signal connected to the trigger input of the ADC, and three control signals for setting system parameters, such as the coincidence window time, output pulse length, and veto time. The performance of the unit was restricted by system clock frequency. The maximum clock frequency was 250 MHz when the unit was implemented to XILINX Spartan 3 FPGA. The unit had a minimum 2 ns of coincidence timing window when it runs in the double data rate (DDR) mode. The unit was also applied to PET image acquisition and the PET/CT and PET/MR images were successfully acquired. The coincidence unit developed in this study was useful for PET image acquisition with PMT- and SiPM-based PET prototypes. The unit can be easily extended to larger systems that have many more detector modules without extra components.
Non-magnetic compliant finger sensor for continuous fine motor movement detection
Springer Science and Business Media LLC - Tập 7 - Trang 215-219 - 2017
A non-magnetic MEG compatible device has been developed that provides continuous force and velocity information. Combined with MEG, this device may find utility in characterizing brain regions associated with force and velocity relative to individual digits or movement pattern. 15 healthy right-handed participants were given visual cues to perform random finger movements on the prototype finger sensor for 21 s and then rest for 21 s (7 times). Respective finger flexion data were obtained, during 151-channel MEG brain scanning, by feeding the signal from finger sensor into four input Analog to Digital Converter (ADC) channels in the MEG hardware. The source activity was reconstructed in beta band using a Linearly Constrained Minimum Variance (LCMV) beamformer in the beta band. The ADC channels were used as regressors for a continuous time General Linear Model (GLM) and a Region of Interest (ROI) was identified to examine activity. MEG analysis showed bilateral activation in the primary motor cortex region. Because individual digits could be isolated in the ADC data, somatotopy of the fingers were observed consistent with the homunculus except pinky finger. The total span was calculated to be 5.5662 mm. The study confirms that the finger sensor is magnetically compatible with MEG measurements and may potentially provide a means to study complex sensorimotor functions. Improved isolation of individual digit information along with the use of machine learning algorithms can help retrieve more accurate results.
Function mapped trajectory estimation for ECG sets
Springer Science and Business Media LLC - Tập 4 Số 3 - Trang 277-284 - 2014
Tái tạo và phân đoạn bề mặt thận cho hướng dẫn phẫu thuật bằng hình ảnh trong phẫu thuật cắt thận nội soi một phần Dịch bởi AI
Springer Science and Business Media LLC - Tập 13 - Trang 165-174 - 2023
Môi trường phẫu thuật động không dự đoán được khiến việc đo lường thông tin hình thái của mô mục tiêu theo thời gian thực trở nên cần thiết cho việc dẫn đường hình ảnh trong phẫu thuật nội soi. Phương pháp thị giác 3D cho tái tạo mô trong phẫu thuật có tiềm năng lớn nhất cho sự phát triển lâm sàng nhờ vào độ chính xác tái tạo cao và tính tương thích với phẫu thuật nội soi. Tuy nhiên, các phương pháp thị giác 3D hiện có gặp khó khăn trong việc đạt được độ chính xác tái tạo cao trong thời gian thực. Ngoài ra, kết quả tái tạo mô trong phẫu thuật thường chứa nhiều thông tin phức tạp từ nền và dụng cụ, điều này cản trở sự phát triển lâm sàng cho các hệ thống dẫn đường hình ảnh. Lấy phẫu thuật cắt thận nội soi một phần (LPN) làm đối tượng nghiên cứu, bài báo này thực hiện việc tái tạo dày đặc theo thời gian thực và tách rời bề mặt mô thận. Thuật toán khớp khối bán toàn cầu dựa trên Census đối xứng trung tâm được đề xuất để tạo ra một bản đồ tương quan dày đặc. Một cơ chế phân đoạn kết nối pixel từng điểm dựa trên GPU được thiết kế để phân đoạn khu vực mô thận. Các phân tích trên tim lợn trong ống nghiệm, thận lợn sống và dữ liệu LPN lâm sàng ngoại tuyến đã được thực hiện để đánh giá độ chính xác và hiệu quả của phương pháp của chúng tôi. Thuật toán đạt được độ chính xác tái tạo là ±2 mm với tỷ lệ cập nhật thời gian thực là 21 fps cho kích thước hình ảnh HD là 960 × 540, và độ chính xác phân đoạn mô mục tiêu đạt 91,0% ngay cả với sự che khuất của dụng cụ phẫu thuật. Kết quả thí nghiệm đã chứng minh rằng phương pháp được đề xuất có thể tái tạo và tách rời bề mặt thận một cách chính xác theo thời gian thực trong phẫu thuật LPN. Kết quả đo lường có thể được sử dụng trực tiếp cho các hệ thống dẫn đường hình ảnh. Phương pháp của chúng tôi cung cấp một cách mới để đo lường thông tin hình học của mô mục tiêu trong quá trình phẫu thuật nội soi.
#phẫu thuật nội soi #tái tạo mô #thị giác 3D #phân đoạn hình ảnh #cắt thận nội soi một phần
Feasibility study of pulse compression technique to improve accuracy of ultrasonic temperature estimation
Springer Science and Business Media LLC - - 2016
Sneaky emotions: impact of data partitions in affective computing experiments with brain-computer interfacing
Springer Science and Business Media LLC - Tập 14 - Trang 103-113 - 2023
Brain-Computer Interfacing (BCI) has shown promise in Machine Learning (ML) for emotion recognition. Unfortunately, how data are partitioned in training/test splits is often overlooked, which makes it difficult to attribute research findings to actual modeling improvements or to partitioning issues. We introduce the “data transfer rate” construct (i.e., how much data of the test samples are seen during training) and use it to examine data partitioning effects under several conditions. As a use case, we consider emotion recognition in videos using electroencephalogram (EEG) signals. Three data splits are considered, each representing a relevant BCI task: subject-independent (affective decoding), video-independent (affective annotation), and time-based (feature extraction). Model performance may change significantly (ranging e.g. from 50% to 90%) depending on how data is partitioned, in classification accuracy. This was evidenced in all experimental conditions tested. Our results show that (1) for affective decoding, it is hard to achieve performance above the baseline case (random classification) unless some data of the test subjects are considered in the training partition; (2) for affective annotation, having data from the same subject in training and test partitions, even though they correspond to different videos, also increases performance; and (3) later signal segments are generally more discriminative, but it is the number of segments (data points) what matters the most. Our findings not only have implications in how brain data are managed, but also in how experimental conditions and results are reported.
Tổng số: 417
- 1
- 2
- 3
- 4
- 5
- 6
- 10