Prizes: Event prize pool from https://ubacm.org/hack/prizes
SO Chat Commander is a Chrome Extension that adds extensions to the StackOverflow and StackExchange Chat. It allows you to ignore users, generate images, quickly reply, load comics, and much more. I started working on this a bit ago, and have updated it with new features and resolved bugs this year.
A model more intelligent, accessible than any Radiologist. A dense convolutional network model for screening and diagnosing Pneumonia and heat mapping air cavities in XRay images using Class Activation Mapping. Accuracy: 54% (Solo development) Deployed over a Flask web app with minimal user interaction. Pytorch; Flask; Google cloud.
Team: Evan Walley, Asher Lieber, Jasjeev Anand, Justin Knauf Project Description: The project is an app that will run on mobile that will enable visitors of the Kallinikeio Municipal Museum (dated back to 800BC) of Athienou, in Cyprus (near Greece) to visualize 3d models of the artifacts in the museum in augmented reality. Project impact: The app would supplement a museum visit by allowing visitors to virtually handle these artifacts without disturbing the pieces themselves - giving them a novel way of experiencing history. The "virtual" artifacts would also have a small description below. Basically each artifact would generate a QR code and the app scans the QR code and you are able to handle these artifacts virtually.
Best Time saver for UB students! This app helps user to find the fastest printing station to print their job. As a bonus functionality, this app also allows user to check list of available computers in UB Lockwood Cybrary. Video Demo: https://drive.google.com/open?id=16tH9RtrdCj6DL9PxwhMZwzFgYT8frIoQ
• first write the python scripts to collect the Data From NYT And Twitter API. • Write the the python Script to crawl the web scraping data from common crawl Repositories • Provision the AWS GLU ETL pipeline and AWS EMR cluster for Map Reduce jobs • Design the interactive Web interface for Visualising the Big Data • Design the AWS Architecture to convert the project in to Production Ready state. github repo: https://github.com/hiteshsantwani/DIC_LAB_2 Member 1: Hitesh Santwani Member 2: Dithya Sridharan Member 3: Kishore Ravisankar
Are you confused to see a clothing brand at a hackathon? Well you should be! Pineapple Posse is a HIP new clothing company that aims to bring the UNEXPECTED to the front lines. Please check out some of our unique designs on our dropbox! P.S. If you would wear our designs please shoot me an email! For DropBox Username: firstname.lastname@example.org PW: UBHACK
This application displays weather information in between two locations provided by Google directions. This helps people to plan their transport based on weather conditions along the way instead of manually searching weather in every location between a source and destination.
This is a Twitter Search Engine which emulates real-world Google search. When something is queried on search engine related tweets appears based on the document rank. Page ranking is provided by Solr. The data set is comprised of tweets on different topics collected for over a month. Trends and Sentiment Analysis is performed on the dataset and Visualization is provided.
There are many articles floating on the internet discussing the President and his cable news addiction. We know Trump often watches several hours of cable news each day, creating a close relationship between The President and News outlet. Our work draws on the relation between Trumps tweet and Social Media Articles. A number of previous works in this area is done by Journalists or Blogger. The task of going through Trumps everyday tweets and manually verifying if it is inspired by the News or not is very time consuming and biased. We have tried to introduce a model that can find the exact instances and also the match percentage where the tweet is being inspired by the news channel. With our study, we are trying to answer - “How often is what Trump tweets is predicted by News Channels?” We calculate the matching score between the news article and tweet, which enables us to have a quantitative analysis amongst the two. The major challenge we faced was that Trump did not use the same wording from News channel. Most of the time he uses sarcastic language to comment about what he saw in the news. So, for a machine, it is difficult to find the similarity between these two contents. There are some current techniques which find the alignment scores between texts, LID is an example. But because of the different language tone used by the two sources, we get a very poor alignment score. Thus, there is a need for a mechanism which does not just consider alignment score. We take into consideration that the actors used in both News and Trumps tweets are mostly the same. For example, the name of the people involved or location or some state laws. Making use of these named entities we find the matches with particular News Article. Using this mechanism, we are also able to eliminate many of the false positives that we were facing earlier based on a normal alignment approach. Team Members: Akshada Bhor Ruturaj Molawade
Ever think an article was too long to read? With this handy Google Chrome extension, simply click a button to give you a shorter summary of the most important points in the article. Development of this project started and ended on 4/27/2019. Created by: Sam Marchant, Himani Dodeja, Liam Orr, and Caroline Hart.
We are an on demand delivery service for UB student. We aim to deliver all range of needs for UB student for everyday needs. From food, grocery, kitchen appliances and classroom needs. Our android app allow students to request what they need for a small fee. These request will be fullfilled in batches by other UB students who wants to make some money while going to shopping themselves. We will first start with Walmart then wegmans. Continue expanding as population demands.
We introduce a novel method for summarization of whiteboard lecture videos using key handwritten content regions. A deep neural network is used for detecting bounding boxes that contain semantically meaningful groups of handwritten content. A neural network embedding is learned, under triplet loss, from the detected regions in order to discriminate between unique handwritten content. The detected regions along with embeddings at every frame of the lecture video are used to extract unique handwritten content across the video which are presented as the video summary. Additionally, a spatiotemporal index is constructed from the video which records the time and location of each individual summary region in the video which can potentially be used for content-based search and navigation. We train and test our methods on the publicly available AccessMath dataset. We use the DetEval scheme to benchmark our summarization by recall of unique ground truth objects and total number of summary regions compared to the ground truth.The most recent version of the publicly available code on some prior work is at the github link below. A powerpoint presentation on some other work on this project is here https://docs.google.com/presentation/d/1vKMSnzR6Wma73jEO-dZ9okELlqNl-ndY3hEwdoXoWn8/edit?usp=sharing The most recent version of this code will be publicly available in the near future.
Biometric security systems such as Apple’s FaceID have replaced passwords for mobile device security, leaving users susceptible to biometric spoofing attacks. In response, Apple has implemented an anti-spoofing scheme, looking for autonomous ocular movement associated within the retina. This research aims to quantify the effectiveness of 3D printed spoof masks in regards to defeating a biometric anti-spoofing system such as Apple’s iPhone X. 3D facial spoofs were created and tested using the iPhone X through 3D printing techniques and live casting. The 3D printed face masks were created from a 3D image utilizing a volumetric regression network. A series of anti-spoofing techniques were used as testing validation. These techniques range from measuring the thermal retention rate, stereoscopic facial image disparity map comparisons between a live face, and a spoof image, and implementation of multitask cascaded convolutional network for recognition of key facial features. Please see more by logging into my DropBox. Username: email@example.com PW: UBHACK