Projects Proposed Projects

Coronary heart disease is one of the diseases with the highest mortality rate. Due to the important position of cardiovascular disease prevention and diagnosis in the medical field, the segmentation of cardiovascular images has gradually become a research hotspot. How to segment accurate blood vessels from coronary angiography videos to assist doctors in making accurate analysis has become the goal of our research. In this project, the student will be required to implement a new method based on the U-net architecture. The proposed method includes three modules: the sequence encoder module, the sequence decoder module, and the sequence filter module. The high-level information of the feature is extracted in the encoder module. Multi-kernel pooling layers suitable for the extraction of blood vessels are added before the decoder module. In the filter block, we add a simple temporal filter to reduce inter-frame flickers.

Noam Rotstein

Asi Elad

Traditional medical imaging techniques (CT, MRI, etc) ironically do not easily enable measurement of surface features. Topographic scans, like those provided by depth cameras, can provide fast objective measurements, but texture maps are typically in the visible spectrum.
Medical screening for scoliosis, inflammation, and skin cancer could potentially benefit from surface scans that include infrared information. Deep learning models that currently rely on RGB input channels could be augmented with a thermal input to improve classification accuracy.
In this project, students will implement a scanning system that combines multiple depth cameras (Azure Kinect) with a thermal camera to produce surface scans with RGB+IR texture.
Project goals:1. Design a calibration target and protocol to register depth/thermal cameras2. Control and record from depth/thermal cameras, in order to automatically calculate and optimize extrinsic camera parameters of an arbitrary number of cameras3. Using calibration parameters, map thermal information onto depth scans
https://doi.org/10.1016/j.procs.2020.05.045   

Benjamin Groisser

In recent years, unmanned aerial vehicles (UAV) are exploited to capture images of large agricultural areas. For agricultural purposes, near infrared (NIR) or other spectral channels are acquired in addition to the RGB colors. These are then used for various tasks, especially semantic segmentation – the ability to distinguish between healthy plants, unwanted weeds, running or standing water, and other classes. We will use an existing semantic segmentation framework, test several network architectures, and employ pre-processing and post-processing algorithms for improved results. Our goal is to participate in an international agriculture segmentation challenge, and attempt to achieve high scores. 

Alon Zvirin

Yaron Honen

Instance segmentation is the task of detecting and masking objects in an image and distinguishing between instances of the same class. Accurate segmentation of leaves in plant images is important in many agricultural applications. These include early detection of water and heat stress, identification of biological infection, monitoring of plant growth, and prediction of harvest yields. Our purpose is to use an existing instance segmentation deep neural network, and integrate it with image processing tools for better performance. In this project we will improve upon previous results, participate in an international leaf segmentation challenge, and attempt to achieve best scores

Alon Zvirin

Yaron Honen

Synthetic cannabinoids are harmful products manufactured and distributed to circumvent legal regulations. Real cannabis can be distinguished from synthetic cannabinoids by inspecting their leaf cystoliths. Cystoliths are hairy appearing outgrowths on leaves and flowers of some plant species. The unaided eye cannot distinguish the real from the fake, but looking through a microscope, it is possible to do so. In this project we will build an artificial intelligence application for classifying real vs. fake cannabis. 

Alon Zvirin

A few years ago, we showed for the first time that the molecular profile of cancer can be predicted by analysis of biopsy images, without using molecular assays. In other words, the shape of the tumor cells and the tissue architecture hold information that allows to accurately predict molecular expression, even though such molecules cannot be seen by humans by visual examination of biopsy images.  Based on these findings, in the past two years we established collaborations with several medical data hospitals in Israel and abroad and extended the scope of our research to different prediction tasks in breast cancer, lung cancer, and Leukemia. We collected and scanned high quality well annotated tens of thousands of histology slides, and extended our research team, consisting of data scientists, graduate and postdoctoral students and clinical collaborators and advisors. We have recently shown how to steer our technology into assisting personalized medicine procedures. Take a part in developing an AI-based framework for analysis of histology images for improving personalized oncology  -  Link to the article 

Gil Shamai

Tal Neoran

Roy Velich

The main task is to adjust a 3D (pre-prepared) CT model on a patient's abdomen and chest, while finding the location and angle at which the model should be allocated, so that it fits exactly to the patient's body position. The project extensively uses voice commands, spatial mapping construction, and receiving and processing data from the sensors on the HoloLens glasses, as well as using an ICP algorithm for the purpose of fitting the model onto the patient. As the major problem the system still misses is the accurate registration of the images delivered by the augmented device (Microsoft Hololens) on the patient’s body. Here is where your knowledge will transform digital work into daily medicine. The student will be part of the development team here in Sheba, coming to the operating room and working with the patients and the surgeons.

Amit Bracha

Yaron Honen

מחלת לב איסכמית נגרמת כתוצאה מהיצרויות וסתימות של כלי דם כליליים. הטיפול כולל תרופות, צנתורים וניתוח מעקפים במידת הצורך.

בצנתור לב מוזרק חומר ניגוד דרך קטטר לעורקים הכליליים תוך כדי שיקוף וכך מופק סרטון בו ניתן לראות הצירויות או חסימות של כלי הדם. במידה ויש צורך בטיפול מועבר תיל דרך ההיצרות ועל גבי התיל מועבר בלון ו/או סטנט לצורך פתיחת ההיצרות.

בארה״ב נעשים כמיליון צנתורים טיפוליים בשנה. כחצי המצנתרים אינם מבצעים את מינימום הצנתורים הנדרש ע״י האיגוד המקצועי ע״מ לשמור על כשירות.

 בכ- 1.1% מהצינתורים מופיע סיבוך של דיסקציה (היפרדות דפנות העורק) של העורק הכלילי כתוצאה מהקטטר, ובכ- 0.35% מהצינתורים פרפורציה של עורק עקב התיל. סיבוכים אלו כרוכים בעליה משמעותית בתמותה ותחלואה של החולים וניתנים למניעה ע״י ביצוע צינתור באופן זהיר ובטיחותי בחלק ניכר מהמקרים.

הסטודנטים יפתחו מערכת עיבוד תמונה המזהה מסרטון הצנתור, בזמן אמת, מצבים מסוכנים שעלולים לגרום לסיבוכים של דיסקציה או פרפורציה של כלי דם כליליים ומתריעה בהתאם, בשיטות של למידת מכונה תוך שימוש במאגר צנתורים של רמב״מ 

Amit Bracha

Yaron Honen

רקע בעיבוד תמונות

נסיון בשיטות למידה