Hello,
I tried to run the application in full deployment (thermEye capture + BLE + android APK) but unfortunately, it was not working properly. Indeed, after few iteration, the application is "Frozen": nothing appears on the logs and the recognition no seems to work.
I think the problem is on the cluster call. Something should not be clear properly and the memory is full.
I solved this problem with a less optimal solution: i just create a _pidevice and a _picluster in the loop of the peopleDetection function. After that, i open the cluster and call the graph constructor (_lynredCNNConstruct function). The rest of the loop does not change except in the end where i destruct the CNN (_lynredCNNDestruct function) and close the cluster (_pi_clusterclose function). With this, the application works ant does not crash.
So i am wondering if anyone have the same problem in a full application deployment. If so, can my solution temporarily works ?
Hello, I tried to run the application in full deployment (thermEye capture + BLE + android APK) but unfortunately, it was not working properly. Indeed, after few iteration, the application is "Frozen": nothing appears on the logs and the recognition no seems to work.
I think the problem is on the cluster call. Something should not be clear properly and the memory is full.
I solved this problem with a less optimal solution: i just create a _pidevice and a _picluster in the loop of the peopleDetection function. After that, i open the cluster and call the graph constructor (_lynredCNNConstruct function). The rest of the loop does not change except in the end where i destruct the CNN (_lynredCNNDestruct function) and close the cluster (_pi_clusterclose function). With this, the application works ant does not crash.
So i am wondering if anyone have the same problem in a full application deployment. If so, can my solution temporarily works ?
Thanks a lot.
NB: I use the GAPoC_b_2.3 with the 3.6.1_dev SDK.