You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+17-17Lines changed: 17 additions & 17 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -1,15 +1,15 @@
1
1
# SynAnno
2
2
3
3
## Live Demo
4
-
A demo is available [here](http://16.170.214.77/)
4
+
A demo is available [here](http://16.170.214.77/reset)
5
5
6
6
## Table of Contents
7
7
8
-
SynAnno is a tool designed for proofreading and correcting synaptic polarity annotation from electron microscopy (EM) volumes - specifically the [H01](https://h01-release.storage.googleapis.com/landing.html) dataset. SynAnno is aimed for integration with CAVE (Connectome Annotation Versioning Engine).
8
+
SynAnno is a tool designed for proofreading and correcting synaptic annotations from electron microscopy (EM) volumes - specifically the [H01](https://h01-release.storage.googleapis.com/landing.html) dataset. SynAnno is aimed for integration with CAVE (Connectome Annotation Versioning Engine).
9
9
10
10
-[Key Components and Subjects](#key-components-and-subjects)
@@ -24,10 +24,10 @@ SynAnno is a tool designed for proofreading and correcting synaptic polarity ann
24
24
-[Navigating SynAnno](#navigating-synanno)
25
25
-[Landing Page](#landing-page)
26
26
-[Open Data](#open-data)
27
-
-[Annotate](#annotate)
28
-
-[Categorize](#categorize)
27
+
-[Error Detection](#error-detection)
28
+
-[Error Categorization](#error-categorization)
29
29
-[Export Annotations](#export-annotations)
30
-
-[Draw](#draw)
30
+
-[Error Correction](#error-correction)
31
31
-[Export Masks](#export-masks)
32
32
-[Contributing](#contributing)
33
33
@@ -37,9 +37,9 @@ SynAnno is a tool designed for proofreading and correcting synaptic polarity ann
37
37
38
38
Harvard's Lichtman laboratory and Google's Connectomics team released the [H01](https://h01-release.storage.googleapis.com/landing.html) dataset, a 1.4 petabyte view of human brain tissue via nanoscale EM. It covers a volume of ~1mm³, featuring tens of thousands of neurons, millions of neuron fragments, 183 million annotated synapses, and 100 proofread cells.
39
39
40
-
### Synaptic Polarity Annotation
40
+
### Direction of information flow
41
41
42
-
Synaptic polarity refers to the directionality of information flow between two neurons at a synapse, the junction where they communicate. In this context, we classify synapses as pre-synaptic (information senders) or post-synaptic (information receivers). A pre-synaptic neuron sends neurotransmitter signals across the synaptic cleft to the post-synaptic neuron, which receives these signals and processes the information. While the segmentation masks highlight the synaptic clefts between neurons, the pre-/post-synaptic IDs are coordinates placed into the associated neurons, identifying the specific sender and receiver. Identifying these key elements is crucial for creating accurate structural and functional neural maps. SynAnno assists in proofreading, correcting, and identifying these segmentation masks and synaptic IDs.
42
+
We classify synapses as pre-synaptic (information senders) or post-synaptic (information receivers). A pre-synaptic neuron sends neurotransmitter signals across the synaptic cleft to the post-synaptic neuron, which receives these signals and processes the information. While the segmentation masks highlight the synaptic clefts between neurons, the pre-/post-synaptic IDs are coordinates placed into the associated neurons, identifying the specific sender and receiver. Identifying these key elements is crucial for creating accurate structural and functional neural maps. SynAnno assists in proofreading, correcting, and identifying these segmentation masks and synaptic IDs.
[CAVE](https://www.biorxiv.org/content/10.1101/2023.07.26.550598v1) is a computational infrastructure to host petabyte connectomes for distributed proofreading and dynamic spatial annotation. Soon, we will provide support to use SynAnno together with CAVE-hosted datasets.
@@ -118,13 +118,13 @@ Repository includes a Dockerfile that enables you to build and run the applicati
118
118
3.**Build the Docker image**: Build the Docker image using the provided Dockerfile.
4.**Run the Docker container**: Run the Docker container using the image you just built.
125
125
126
126
```bash
127
-
docker run -d --name synanno -p 80:80 -p 9015:9015 synanno
127
+
docker run -d --name synanno -p 80:80 -p 9015:9015 synanno/uwsgi
128
128
```
129
129
130
130
5.**Access the application**: Open a web browser and go to `http://localhost` to access the application running in the Docker container.
@@ -230,7 +230,7 @@ python run.py
230
230
231
231
- URL: http://127.0.0.1:5000/
232
232
233
-
On the landing page you can choose between two workflows: "Revise Dataset" and "Proofread Annotation". The former allows you to redraw segmentation masks, assign pre-/post-synaptic IDs, and add missed false negatives. The latter allows you to review existing segmentations and pre-/post-synaptic IDs, mark incorrect instances, and assign error descriptions to those instances. After categorizing all erroneous instances in the "Proofread Annotation" workflow, you can automatically proceed with the "Revise Dataset" workflow.
233
+
On the landing page you can choose between two workflows: "Error Correction" and "Error Detection". The former allows you to redraw segmentation masks, assign pre-/post-synaptic IDs, and add missed false negatives. The latter allows you to review existing segmentations and pre-/post-synaptic IDs, mark incorrect instances, and assign error descriptions to those instances. After categorizing all erroneous instances in the "Error Detection" workflow, you can automatically proceed with the "Error Correction" workflow.
234
234
235
235
Each page has three buttons: "Home", "Question Mark", and "Menu". The first returns you to the landing page. The second provides an explanation of the current view and its functionality. The third provides general information and contact details.
236
236
@@ -240,7 +240,7 @@ Each page has three buttons: "Home", "Question Mark", and "Menu". The first retu
240
240
241
241
- URL: http://127.0.0.1:5000/open_data
242
242
243
-
This view is identical forboth workflows. You'll be prompted to provide a source bucket and target bucket (bothin Neuroglancer's precomputed format), a URL to a materialization table, bucket secrets (defaults to ~/.cloudvolume/secrets), and optionally a JSON file containing instance metadata. The JSON file can be used to save and restore sessions or start a "Revise Dataset" workflow with information from a "Proofread Annotation" workflow. If you wish to use "View-Centric" neuron selection in a "Proofread Annotation" workflow, you must also provide a neuropil segmentation bucket in Neuroglancer's precomputed format as well.
243
+
This view is identical forboth workflows. You'll be prompted to provide a source bucket and target bucket (bothin Neuroglancer's precomputed format), a URL to a materialization table, bucket secrets (defaults to ~/.cloudvolume/secrets), and optionally a JSON file containing instance metadata. The JSON file can be used to save and restore sessions or start a "Error Correction" workflow with information from a "Error Detection" workflow. If you wish to use "View-Centric" neuron selection in a "Error Detection" workflow, you must also provide a neuropil segmentation bucket in Neuroglancer's precomputed format as well.
244
244
245
245
Opening the "Volume Parameters" tab, you will have two different options forthe mannerin which to selectyour instances, "View Centric" and "Neuron Centric".
246
246
@@ -258,12 +258,12 @@ If you choose the 'Neuron Centric' approach. You'll need to specify the coordina
258
258
259
259
After providing the required information, click 'Submit' to prepare the data for the first page or revision. Then, click "Start Data Proofread"/"Start Drawing" to begin proofreading or revision.
260
260
261
-
### Annotate
261
+
### Error Detection
262
262
263
263
- Workflow: Proofreading
264
264
- URL: http://127.0.0.1:5000/annotation
265
265
266
-
Clicking `Start Data Proofread` directs you to a grid view of instances. Instance status is indicated by color: correct (green), incorrect (red), unsure (gray). Clicking on an instance changes its status: once for incorrect, twice for unsure, three times for correct.
266
+
Clicking `Start Data Proofread` directs you to a grid view of instances. Instance status is indicated by color: correct (green), incorrect (red), unsure (yellow). Clicking on an instance changes its status: once for incorrect, twice for unsure, three times for correct.
267
267
268
268
[![Grid View][3]][3]
269
269
@@ -277,7 +277,7 @@ Click `View in NG` to view the instance in Neuroglancer.
277
277
278
278
After evaluating the segmentation masks, click `->` to load and evaluate the page. When done, click `Error Processing` on the last proofreading page.
279
279
280
-
### Categorize
280
+
### Error Categorization
281
281
282
282
- Workflow: Proofreading
283
283
- URL: http://127.0.0.1:5000/categorize
@@ -295,11 +295,11 @@ If you marked instances as false positives, you'll be asked if they should be di
295
295
- Workflow: Proofreading
296
296
- URL: http://127.0.0.1:5000/export_annotate
297
297
298
-
After clicking `Submit and finish`, you can download the JSON file containing instance metadata by clicking `Download JSON`, redraw masks with the `Revise Dataset` workflow by clicking `Redraw Masks`, or start a new process by clicking `Start New Process`.
298
+
After clicking `Submit and finish`, you can download the JSON file containing instance metadata by clicking `Download JSON`, redraw masks with the `Error Correction` workflow by clicking `Redraw Masks`, or start a new process by clicking `Start New Process`.
0 commit comments