WIP: refs #analysis updates on analysis commit

analysis/server/24-02-2021 - Findings & Updates.txt

@@ -0,0 +1,43 @@
+
+3 Phases,
+
+Phase 1,
+
+Automated/ Semi-automated Annotation Platform,
+- Can upload image or CCTV videos
+- Search bar linked with Google image API and Bing Image API
+- Training Cloud - Inference hardware (T4's)
+- Uploaded image go into "Training Cloud" for training "selection models"
+- Reference - TrainingSet.AI video link
+
+
+Phase 2,
+
+Training cloud,
+
+- Same as Inference hardware (T4's)
+
+EC2 URL [VISTA Server as used by G-SATE application] - https://ec2-54-152-186-179.compute-1.amazonaws.com
+
+What is T4 Inference hardware?
+Where to get TrainingSet.AI video link?
+
+Where to get tCloudBrowserExtension_VERSION_NUMBER. zip?
+
+=====================================================
+
+Training cloud --
+
+For training, 250 for positive dataset and 50 for validation dataset
+tCloud training interface - http://training.graymatics.com [Not Available]
+
+=====================================================
+
+Working with the application,
+
+1. Start the client and server using "npm start" command.
+2. Access the application using the URL - http://localhost:4200/objectDetection
+3. Create a folder named "uploads" in server folder.
+4. Use "Browse" button to attach the image and upload the same using "Upload Image" button [Only JPG, JPEG and PNG files allowed]
+
+

analysis/server/25-02-2021 - Findings & Updates.txt

@@ -0,0 +1,33 @@
+
+As per our understanding, the following points are given
+
+1. The architecture of G-SATE application is like given below,
+
+	Front-End [Angular - G-SATE]	+	Back-End [Node Express JS - G-SATE + Django AI - VISTA]		+	MySQL DB
+
+2. After refering the video links of TrainingSet.AI, we infer the following understandings,
+
+		A. Gets images from the internet using GET IMAGES FROM THE INTERNET button in the annotation creation screen or using AWS S3 storage
+		B. Roles and responsibilities in TrainingSet.AI,
+			-> Client -	The person who creates the "Annotation Tasks" and uploads the images and corresponding labels for them to be identified and assign them to the 							"Annotator".
+					Ex: Upload the image of street traffic with labels such as traffic lights, cars, bikes, cycles, pedestrians, etc., to be annotated in them.
+			-> Annotator -	The person who views the Annotation Tasks and annotate the objects in the image uploaded by the client.
+					Ex: Annotate the objects with the user defined labels such as cars, bikes, etc., using shapes such as square, rectangle, polygon, etc.
+
+			-> QA -		The person who is responsible to view the Annotation Tasks with the identified objects and can perform Accept or Reject action on the task.
+					Ex: Task can be accepted if all the objects are annotated correctly or can reject if missing or incorrect annotations.
+
+
+3. Queries from mail conversation subjected as "Graymatics - TrainingSet.AI integration",
+
+	>> We did have developed an annotation platform but it might not be as robust as TrainingSet.ai
+	Is the developed annotation platform by GrayMatics is refering to VISTA server?
+
+	>> While we could use their platform, we may need to expand the capabilities to include the various additional entities
+	As of now, is the G-SATE application uses the TrainingSet.AI platform for object annotation of the images?
+
+	>> Build on our platform internally to bring in the pertinent features from TrainingSet.AI
+	Is "our platform" refering to the VISTA server and is it extending direct features from TrainingSet.AI?
+
+	>> Towards this, we can also bring in an outsourced front-end s/w services company that could develop this with our guidance
+	Is the term, "front-end s/w services" refering to the application G-SATE?

analysis/server/26-02-2021 - Findings & Updates.txt

@@ -0,0 +1,14 @@
+
+On Discussing with Dhananjay and Pramod, the following were concluded,
+
+1. Reference of TrainingSet.AI for the deveployment of G-SATE application.
+2. Brief introduction of VISTA server and it's features
+3. Area to be focused specifically in the application [Object Detection]
+4. Bugs and specification of the existing G-SATE applications
+5. Installation and setup guidelines to use tCloudBrowserExtension
+
+Upcoming proceedings,
+
+1. To show the annotation boxes visually in the canvas once the image is uploaded for annotation
+2. To create new annotations with user defined labels
+3. To get the tCloudBrowserExtension from Lusong and integrated the same into the G-SATE application