Cyber-Security environment has always been characterized by a continuous process, in which every technique used to detect threats was shortly followed by a malware adjustment meant to evade that detection approach. While machine learning and emulators have been an important part of this process for several years now, translating features extracted from malware samples into images and feeding them as input to convolutional neural networks is a more recent endeavor in this area. This paper presents our research on several methods that can be used in order to translate features extracted from malware emulation during static analysis (with a special focus on features that reflect anti-emulation techniques) for threat detection.