FedCAM - Identifying Malicious Models in Federated Learning Environments Conditionally to Their Activation Maps | IEEE Conference Publication | IEEE Xplore