Top Rated Segments.ai Alternatives
I love all its features and how easy it is for both 2D and 3D labelling. The dataset creation and export is simple and intuitive. Labelling, especially in 2D, relies on many tools (superpixels, brushes, etc) that make the experience easy and straightforward to explain to an inexperienced person. Review collected by and hosted on G2.com.
The 3D visualization gets stuck at times because it's very computationally heavy. It would be cool to introduce a feature for "hierarchical labeling", meaning to have multiple instance ids for the same pixel/point. This is useful, for example, in the agricultural domain where both plant instance and leaf instance segmentation are useful, thus each pixel has to be associated to a specific plant and to a specific leaf. Review collected by and hosted on G2.com.
17 out of 18 Total Reviews for Segments.ai
I am a master's student, and I couldn't find a good tool for 3D point cloud annotation until I found Segments.ai. They generously provided me with a free academic license. This tool was exactly what I needed. It is easy to use, accepts general formats like PCD (not just BIN formats), and they send helpful emails and Google Colab notebooks. Review collected by and hosted on G2.com.
There isn't much to dislike about Segments.ai. I wish they made more videos on YouTube as they are very helpful. Review collected by and hosted on G2.com.
The best thing about segments.ai is how fast one can label images. With other tools that I have used it could take up to 5 minutes to label one image. But with segments.ai it can take as little as 10-20 seconds to label one image with several classes.
The reason for why it is so easy to label images with segments.ai is that this product comes with several different features such as edit with superpixels, brush, polygon etc and it also comes with autosegmentation (which as of now is in beta mode).
Moreover, as a manager it is also very easy to se statistics for each labeler and this product also has a very nice review function so that you easily can correct mistakes. However, mistake seldomly happens, since this tool make it so easy to label iamges.
Furthermore, once you are done with a labeling project there is an amazing python SDK that makes it very easy to download both images and segmentation bitmaps.
I would also like to add that when I encounter a problems with segments.ai (which rarely happens), their team are very quick to respond and they are extremely helpful. Review collected by and hosted on G2.com.
So far I have not encountered anything that I dislike about segments.ai Review collected by and hosted on G2.com.
An easy interface and export function. Free academic access. Review collected by and hosted on G2.com.
It would be nice to have options to export annotations straight from the web version, without using Pythin SDA. Review collected by and hosted on G2.com.
Versatile Annotation Options: The diverse range of annotation types, from bounding boxes to keypoints, empowers me to accurately label various types of data, ensuring comprehensive annotation for my projects.
AI-Assisted Labeling: The AI-powered pre-labeling feature is a true time-saver. It not only speeds up the annotation process but also enhances accuracy by intelligently predicting labels, elevating the quality of my annotated datasets. Review collected by and hosted on G2.com.
Complexity for New Users: Data annotation tools can sometimes have a learning curve, especially for users who are new to the concept of data labeling or have limited technical expertise. Navigating the platform, understanding annotation types, and using advanced features might be challenging initially.
Quality Control: Ensuring the accuracy and consistency of annotations can be a challenge, especially when dealing with complex or large datasets. Users might face difficulties in maintaining high-quality annotations across different annotators or addressing discrepancies in labeling. Review collected by and hosted on G2.com.
Segments.ai is very accurate, even for complex images. Also Segments.ai is very fast, even for large images. The software can process images quickly and efficiently. Review collected by and hosted on G2.com.
Segments.ai can have a steep learning curve, especially for users who are not familiar with image annotation software. The documentation for Segments.ai is a little lacking. For example, there is no documentation on how to use the software for machine learning tasks. Review collected by and hosted on G2.com.
With the help of Segments.ai you can easily label your images with appropriate class. As we know that labelling is a time consuming process and Segments.ai solves this problem very easily. It has an amazing python SDK which makes your life too easy by providing the facility of downloading the images and segmentation bitmaps. Their support team is very helpful and quick with response so that you can smoothly done your work. Review collected by and hosted on G2.com.
I think there is nothing what I dislike about Segments.ai. It us bery helpful the only thing they have to imprive is that please ad some more documentation data. Review collected by and hosted on G2.com.
I like that the software provides different levels of coarseness for the segmentation, helping the user to catch small details or big objects at once without the need to restart it. It is also easy to correct the segments whenever small parts are wrongly assigned to them, using the coarseness level or just doing it by hand. Review collected by and hosted on G2.com.
I think there is still room for improvement when segmenting 3D inputs, especially large ones that sometimes make the software slow. But considering the difficulty of labeling such type of data, segments.ai is still very useful and a promising way to reduce the required effort. Review collected by and hosted on G2.com.
The ease of using the editor to label my dataset, as dataset creator, segment.ai was a super helpful to label my dataset for panoptic segmentation Review collected by and hosted on G2.com.
maybe when you have multiple objects, switching between labeling the objects is a bit hard Review collected by and hosted on G2.com.
My team was on the lookout for a tool to help us verify our 3D data algorithms, and was excited to see that segments.ai had a feature to overlay 3D point clouds with 2D images. This proved very useful at giving labellers context for what they were looking at in 3D, especially in noisy shots.
Another feature that was very useful for us was the ability to visualise intensity point cloud, which was missing from several other platforms that we researched. Being able to visualise the intensity data well was critical for interpreting our point cloud meaningfully. Segments.ai a range of nice visualisation options such as point size and colour pallets, for us to toggle as we liked.
Finally, the 3D label features were very handy. The configuration option to generate bounding boxes at pre-set object sizes was useful for us as we were working with a single object of known size. The interpolation between frames for point cloud sequences was also great, auto-generating estimated bounding boxes for subsequent frames after the first manual label. These features saved us a lot of time.
Excellent labelling platform, especially for 3D data. Review collected by and hosted on G2.com.
Some features had a handful of development holes, such as options missing from configs or just missing from documentation. Also a handful of bugs were noticed by my team, but the segments.ai team were extremely responsive at fixing these when they were reported. Overall the platform seemed at the level of development that things aren't always perfect, but problems resolved themselves quickly enough to not pose a major issue. Review collected by and hosted on G2.com.
