Open happyprotean opened 5 years ago
Hi sir, I use your model to detect baby cry and the test accuracy and F1 score are both around 0.95. But the question is the performance is too great. I come up two reasons to explain this. First, the test dataset is too simple. Second, I notice that some of the data examples you provided come from the same audio which means long aduio is cut to several 5 seconds clips. Maybe this contributes to the test performance. And I am working on detect baby cry recently but find there is no public big dataset for this goal. I collect some data here:ESC-50, donateacry,DCASE-2017 rare event detection and this respository. Do you know some other baby cry dataset? It would be much appreciated if you let me know.
SPLANN is a project for baby crying .but I don't how to access it.I come from guangzhou
@liupeng678 Yes, I have look into the SPLANN project page but found no access to this dataset. What a pity!!! Do you also work on baby crying detection?
@liupeng678 Yes, I have look into the SPLANN project page but found no access to this dataset. What a pity!!! Do you also work on baby crying detection?
My previous work was computer vision, and I began to learn baby crying recently. Can you give me a wechat Id or QQ? Maybe we can do it more qucikly.
@liupeng678 Hi you can add my wechat: fhx456321
Have you got SPLANN, can I have the download link, I am also conducting a research related to baby cry
@cuong3004 did you manage to get the link? I would also be interested. Thank you very much!
Hello there, I am also conducting a research related to baby cry so i need a data for this project. Any have SPLANN for this.
Hello, can I have too the download link, I am also conducting a project related to baby cry.
Hello there, can I have the download link please, I am also conducting a project related to baby crying.
Hello, I am also doing a project related to baby cry, but there are not a lot of datasets available for this. I could only find donateacry. Can anyone help me with finding another dataset
Hello , can I have the download link please, I am also conducting a project related to baby cry.
Hello, as a grad student working on the infant cry project, I also appreciate the link of access.
Hi sir, I use your model to detect baby cry and the test accuracy and F1 score are both around 0.95. But the question is the performance is too great. I come up two reasons to explain this. First, the test dataset is too simple. Second, I notice that some of the data examples you provided come from the same audio which means long aduio is cut to several 5 seconds clips. Maybe this contributes to the test performance. And I am working on detect baby cry recently but find there is no public big dataset for this goal. I collect some data here:ESC-50, donateacry,DCASE-2017 rare event detection and this respository. Do you know some other baby cry dataset? It would be much appreciated if you let me know.